diff --git a/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_en.md b/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_en.md new file mode 100644 index 00000000000000..54ec549d0a0d8b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English albaniani_sum_v1 T5Transformer from mHossain +author: John Snow Labs +name: albaniani_sum_v1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`albaniani_sum_v1` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/albaniani_sum_v1_en_5.4.2_3.0_1723915533205.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/albaniani_sum_v1_en_5.4.2_3.0_1723915533205.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("albaniani_sum_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("albaniani_sum_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|albaniani_sum_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|947.2 MB| + +## References + +https://huggingface.co/mHossain/Albaniani_sum_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_pipeline_en.md new file mode 100644 index 00000000000000..70db47795e7564 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-albaniani_sum_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English albaniani_sum_v1_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: albaniani_sum_v1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`albaniani_sum_v1_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/albaniani_sum_v1_pipeline_en_5.4.2_3.0_1723915596902.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/albaniani_sum_v1_pipeline_en_5.4.2_3.0_1723915596902.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("albaniani_sum_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("albaniani_sum_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|albaniani_sum_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|947.2 MB| + +## References + +https://huggingface.co/mHossain/Albaniani_sum_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_en.md b/docs/_posts/ahmedlone127/2024-08-17-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_en.md new file mode 100644 index 00000000000000..e105daefb34dfd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English all_mt5_base_10_spider_15_wikisql_nepal_bhasa T5Transformer from e22vvb +author: John Snow Labs +name: all_mt5_base_10_spider_15_wikisql_nepal_bhasa +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_mt5_base_10_spider_15_wikisql_nepal_bhasa` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_mt5_base_10_spider_15_wikisql_nepal_bhasa_en_5.4.2_3.0_1723939059076.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_mt5_base_10_spider_15_wikisql_nepal_bhasa_en_5.4.2_3.0_1723939059076.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("all_mt5_base_10_spider_15_wikisql_nepal_bhasa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("all_mt5_base_10_spider_15_wikisql_nepal_bhasa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_mt5_base_10_spider_15_wikisql_nepal_bhasa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/e22vvb/ALL_mt5-base_10_spider_15_wikiSQL_new \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_en.md b/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_en.md new file mode 100644 index 00000000000000..337a379b663f6b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English attributes_latest T5Transformer from bitadin +author: John Snow Labs +name: attributes_latest +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`attributes_latest` is a English model originally trained by bitadin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/attributes_latest_en_5.4.2_3.0_1723938821896.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/attributes_latest_en_5.4.2_3.0_1723938821896.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("attributes_latest","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("attributes_latest", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|attributes_latest| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bitadin/attributes-latest \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_pipeline_en.md new file mode 100644 index 00000000000000..3e5e2e19dcde25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-attributes_latest_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English attributes_latest_pipeline pipeline T5Transformer from bitadin +author: John Snow Labs +name: attributes_latest_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`attributes_latest_pipeline` is a English model originally trained by bitadin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/attributes_latest_pipeline_en_5.4.2_3.0_1723938871107.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/attributes_latest_pipeline_en_5.4.2_3.0_1723938871107.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("attributes_latest_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("attributes_latest_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|attributes_latest_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bitadin/attributes-latest + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-baguette_en.md b/docs/_posts/ahmedlone127/2024-08-17-baguette_en.md new file mode 100644 index 00000000000000..8d42a895489321 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-baguette_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English baguette T5Transformer from marijacet +author: John Snow Labs +name: baguette +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`baguette` is a English model originally trained by marijacet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/baguette_en_5.4.2_3.0_1723938282548.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/baguette_en_5.4.2_3.0_1723938282548.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("baguette","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("baguette", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|baguette| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/marijacet/baguette \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-baguette_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-baguette_pipeline_en.md new file mode 100644 index 00000000000000..6ecb6f35ac0c6c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-baguette_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English baguette_pipeline pipeline T5Transformer from marijacet +author: John Snow Labs +name: baguette_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`baguette_pipeline` is a English model originally trained by marijacet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/baguette_pipeline_en_5.4.2_3.0_1723938299636.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/baguette_pipeline_en_5.4.2_3.0_1723938299636.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("baguette_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("baguette_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|baguette_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/marijacet/baguette + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_en.md b/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_en.md new file mode 100644 index 00000000000000..15bc4c96d94c51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English banglat5_headline_trial_with_spanish_batch16_2e_4_ep10 T5Transformer from mdosama39 +author: John Snow Labs +name: banglat5_headline_trial_with_spanish_batch16_2e_4_ep10 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5_headline_trial_with_spanish_batch16_2e_4_ep10` is a English model originally trained by mdosama39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_en_5.4.2_3.0_1723916671858.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_en_5.4.2_3.0_1723916671858.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("banglat5_headline_trial_with_spanish_batch16_2e_4_ep10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("banglat5_headline_trial_with_spanish_batch16_2e_4_ep10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5_headline_trial_with_spanish_batch16_2e_4_ep10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.9 MB| + +## References + +https://huggingface.co/mdosama39/banglat5-headline-trial-with-ES-batch16-2e-4-ep10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline_en.md new file mode 100644 index 00000000000000..25ef8b78e21188 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline pipeline T5Transformer from mdosama39 +author: John Snow Labs +name: banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline` is a English model originally trained by mdosama39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline_en_5.4.2_3.0_1723916722069.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline_en_5.4.2_3.0_1723916722069.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5_headline_trial_with_spanish_batch16_2e_4_ep10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.9 MB| + +## References + +https://huggingface.co/mdosama39/banglat5-headline-trial-with-ES-batch16-2e-4-ep10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_en.md b/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_en.md new file mode 100644 index 00000000000000..92c4c5fdaf53cc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bikes_select_t5_small_18 T5Transformer from neal61 +author: John Snow Labs +name: bikes_select_t5_small_18 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_select_t5_small_18` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_select_t5_small_18_en_5.4.2_3.0_1723915496201.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_select_t5_small_18_en_5.4.2_3.0_1723915496201.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bikes_select_t5_small_18","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bikes_select_t5_small_18", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_select_t5_small_18| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.9 MB| + +## References + +https://huggingface.co/neal61/bikes-select-t5-small-18 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_pipeline_en.md new file mode 100644 index 00000000000000..460431bb95a44e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-bikes_select_t5_small_18_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bikes_select_t5_small_18_pipeline pipeline T5Transformer from neal61 +author: John Snow Labs +name: bikes_select_t5_small_18_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_select_t5_small_18_pipeline` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_select_t5_small_18_pipeline_en_5.4.2_3.0_1723915514085.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_select_t5_small_18_pipeline_en_5.4.2_3.0_1723915514085.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bikes_select_t5_small_18_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bikes_select_t5_small_18_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_select_t5_small_18_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.9 MB| + +## References + +https://huggingface.co/neal61/bikes-select-t5-small-18 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_en.md b/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_en.md new file mode 100644 index 00000000000000..542e1a68462023 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bikes_t5_small_9 T5Transformer from neal61 +author: John Snow Labs +name: bikes_t5_small_9 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small_9` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_9_en_5.4.2_3.0_1723913118021.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_9_en_5.4.2_3.0_1723913118021.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bikes_t5_small_9","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bikes_t5_small_9", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small_9| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.5 MB| + +## References + +https://huggingface.co/neal61/bikes-t5-small-9 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_pipeline_en.md new file mode 100644 index 00000000000000..71500f1eb07598 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-bikes_t5_small_9_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bikes_t5_small_9_pipeline pipeline T5Transformer from neal61 +author: John Snow Labs +name: bikes_t5_small_9_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small_9_pipeline` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_9_pipeline_en_5.4.2_3.0_1723913135787.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_9_pipeline_en_5.4.2_3.0_1723913135787.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bikes_t5_small_9_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bikes_t5_small_9_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small_9_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.5 MB| + +## References + +https://huggingface.co/neal61/bikes-t5-small-9 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_en.md b/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_en.md new file mode 100644 index 00000000000000..7bd2f23cb1242a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English boss_sentiment_t5_large T5Transformer from Kyle1668 +author: John Snow Labs +name: boss_sentiment_t5_large +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boss_sentiment_t5_large` is a English model originally trained by Kyle1668. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boss_sentiment_t5_large_en_5.4.2_3.0_1723938294315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boss_sentiment_t5_large_en_5.4.2_3.0_1723938294315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("boss_sentiment_t5_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("boss_sentiment_t5_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boss_sentiment_t5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Kyle1668/boss-sentiment-t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_pipeline_en.md new file mode 100644 index 00000000000000..8ddb652cf82c4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-boss_sentiment_t5_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English boss_sentiment_t5_large_pipeline pipeline T5Transformer from Kyle1668 +author: John Snow Labs +name: boss_sentiment_t5_large_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boss_sentiment_t5_large_pipeline` is a English model originally trained by Kyle1668. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boss_sentiment_t5_large_pipeline_en_5.4.2_3.0_1723938438742.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boss_sentiment_t5_large_pipeline_en_5.4.2_3.0_1723938438742.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("boss_sentiment_t5_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("boss_sentiment_t5_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boss_sentiment_t5_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Kyle1668/boss-sentiment-t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_en.md new file mode 100644 index 00000000000000..789398f4e272cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_manyee101 T5Transformer from Manyee101 +author: John Snow Labs +name: burmese_awesome_billsum_model_manyee101 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_manyee101` is a English model originally trained by Manyee101. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_manyee101_en_5.4.2_3.0_1723936586687.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_manyee101_en_5.4.2_3.0_1723936586687.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_manyee101","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_manyee101", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_manyee101| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/Manyee101/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_pipeline_en.md new file mode 100644 index 00000000000000..e963d5fdbf89e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_manyee101_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_manyee101_pipeline pipeline T5Transformer from Manyee101 +author: John Snow Labs +name: burmese_awesome_billsum_model_manyee101_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_manyee101_pipeline` is a English model originally trained by Manyee101. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_manyee101_pipeline_en_5.4.2_3.0_1723936607392.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_manyee101_pipeline_en_5.4.2_3.0_1723936607392.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_manyee101_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_manyee101_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_manyee101_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/Manyee101/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_en.md new file mode 100644 index 00000000000000..cb1e1c9a5e8572 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_trisert T5Transformer from Trisert +author: John Snow Labs +name: burmese_awesome_billsum_model_trisert +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_trisert` is a English model originally trained by Trisert. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_trisert_en_5.4.2_3.0_1723916362748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_trisert_en_5.4.2_3.0_1723916362748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_trisert","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_trisert", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_trisert| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|315.0 MB| + +## References + +https://huggingface.co/Trisert/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_pipeline_en.md new file mode 100644 index 00000000000000..dc03ccd893c3f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_billsum_model_trisert_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_trisert_pipeline pipeline T5Transformer from Trisert +author: John Snow Labs +name: burmese_awesome_billsum_model_trisert_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_trisert_pipeline` is a English model originally trained by Trisert. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_trisert_pipeline_en_5.4.2_3.0_1723916385235.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_trisert_pipeline_en_5.4.2_3.0_1723916385235.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_trisert_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_trisert_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_trisert_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|315.0 MB| + +## References + +https://huggingface.co/Trisert/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_en.md new file mode 100644 index 00000000000000..49443351b4d8fd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_ahmedssoliman T5Transformer from AhmedSSoliman +author: John Snow Labs +name: burmese_awesome_opus_books_model_ahmedssoliman +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_ahmedssoliman` is a English model originally trained by AhmedSSoliman. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ahmedssoliman_en_5.4.2_3.0_1723926375404.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ahmedssoliman_en_5.4.2_3.0_1723926375404.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_ahmedssoliman","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_ahmedssoliman", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_ahmedssoliman| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/AhmedSSoliman/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_pipeline_en.md new file mode 100644 index 00000000000000..dd1543c714d205 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_ahmedssoliman_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_ahmedssoliman_pipeline pipeline T5Transformer from AhmedSSoliman +author: John Snow Labs +name: burmese_awesome_opus_books_model_ahmedssoliman_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_ahmedssoliman_pipeline` is a English model originally trained by AhmedSSoliman. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ahmedssoliman_pipeline_en_5.4.2_3.0_1723926393857.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ahmedssoliman_pipeline_en_5.4.2_3.0_1723926393857.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_ahmedssoliman_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_ahmedssoliman_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_ahmedssoliman_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/AhmedSSoliman/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_en.md new file mode 100644 index 00000000000000..bee91c33616c23 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_basso4 T5Transformer from basso4 +author: John Snow Labs +name: burmese_awesome_opus_books_model_basso4 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_basso4` is a English model originally trained by basso4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_basso4_en_5.4.2_3.0_1723920753542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_basso4_en_5.4.2_3.0_1723920753542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_basso4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_basso4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_basso4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.5 MB| + +## References + +https://huggingface.co/basso4/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_pipeline_en.md new file mode 100644 index 00000000000000..47cd48a798b822 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_basso4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_basso4_pipeline pipeline T5Transformer from basso4 +author: John Snow Labs +name: burmese_awesome_opus_books_model_basso4_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_basso4_pipeline` is a English model originally trained by basso4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_basso4_pipeline_en_5.4.2_3.0_1723920772497.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_basso4_pipeline_en_5.4.2_3.0_1723920772497.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_basso4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_basso4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_basso4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.5 MB| + +## References + +https://huggingface.co/basso4/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_en.md new file mode 100644 index 00000000000000..9b2888867be0ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_elrosech T5Transformer from elrosech +author: John Snow Labs +name: burmese_awesome_opus_books_model_elrosech +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_elrosech` is a English model originally trained by elrosech. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_elrosech_en_5.4.2_3.0_1723932011231.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_elrosech_en_5.4.2_3.0_1723932011231.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_elrosech","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_elrosech", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_elrosech| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/elrosech/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_pipeline_en.md new file mode 100644 index 00000000000000..a3a0cd80179e14 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_elrosech_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_elrosech_pipeline pipeline T5Transformer from elrosech +author: John Snow Labs +name: burmese_awesome_opus_books_model_elrosech_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_elrosech_pipeline` is a English model originally trained by elrosech. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_elrosech_pipeline_en_5.4.2_3.0_1723932032630.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_elrosech_pipeline_en_5.4.2_3.0_1723932032630.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_elrosech_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_elrosech_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_elrosech_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/elrosech/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_en.md new file mode 100644 index 00000000000000..9a468fb4a480e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_jlau0019 T5Transformer from jlau0019 +author: John Snow Labs +name: burmese_awesome_opus_books_model_jlau0019 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_jlau0019` is a English model originally trained by jlau0019. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_jlau0019_en_5.4.2_3.0_1723918175596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_jlau0019_en_5.4.2_3.0_1723918175596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_jlau0019","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_jlau0019", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_jlau0019| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.4 MB| + +## References + +https://huggingface.co/jlau0019/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_pipeline_en.md new file mode 100644 index 00000000000000..a741fe0bfbb7f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_awesome_opus_books_model_jlau0019_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_jlau0019_pipeline pipeline T5Transformer from jlau0019 +author: John Snow Labs +name: burmese_awesome_opus_books_model_jlau0019_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_jlau0019_pipeline` is a English model originally trained by jlau0019. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_jlau0019_pipeline_en_5.4.2_3.0_1723918193955.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_jlau0019_pipeline_en_5.4.2_3.0_1723918193955.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_jlau0019_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_jlau0019_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_jlau0019_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.4 MB| + +## References + +https://huggingface.co/jlau0019/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_en.md new file mode 100644 index 00000000000000..946b804e45df92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_ielts T5Transformer from duwuonline +author: John Snow Labs +name: burmese_ielts +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_ielts` is a English model originally trained by duwuonline. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_ielts_en_5.4.2_3.0_1723938194688.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_ielts_en_5.4.2_3.0_1723938194688.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_ielts","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_ielts", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_ielts| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/duwuonline/my-ielts \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_pipeline_en.md new file mode 100644 index 00000000000000..530aff60377c06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-burmese_ielts_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_ielts_pipeline pipeline T5Transformer from duwuonline +author: John Snow Labs +name: burmese_ielts_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_ielts_pipeline` is a English model originally trained by duwuonline. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_ielts_pipeline_en_5.4.2_3.0_1723938245025.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_ielts_pipeline_en_5.4.2_3.0_1723938245025.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_ielts_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_ielts_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_ielts_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/duwuonline/my-ielts + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_en.md b/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_en.md new file mode 100644 index 00000000000000..5f272463d9ecb1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cat_sum_french T5Transformer from homersimpson +author: John Snow Labs +name: cat_sum_french +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cat_sum_french` is a English model originally trained by homersimpson. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cat_sum_french_en_5.4.2_3.0_1723926118236.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cat_sum_french_en_5.4.2_3.0_1723926118236.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cat_sum_french","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cat_sum_french", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cat_sum_french| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|989.5 MB| + +## References + +https://huggingface.co/homersimpson/cat-sum-fr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_pipeline_en.md new file mode 100644 index 00000000000000..dfa25c75ad9a13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cat_sum_french_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cat_sum_french_pipeline pipeline T5Transformer from homersimpson +author: John Snow Labs +name: cat_sum_french_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cat_sum_french_pipeline` is a English model originally trained by homersimpson. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cat_sum_french_pipeline_en_5.4.2_3.0_1723926168006.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cat_sum_french_pipeline_en_5.4.2_3.0_1723926168006.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cat_sum_french_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cat_sum_french_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cat_sum_french_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|989.5 MB| + +## References + +https://huggingface.co/homersimpson/cat-sum-fr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_en.md b/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_en.md new file mode 100644 index 00000000000000..2317bac0de327e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English catt5_solr_finetunned2 T5Transformer from oooriii +author: John Snow Labs +name: catt5_solr_finetunned2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`catt5_solr_finetunned2` is a English model originally trained by oooriii. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned2_en_5.4.2_3.0_1723923064821.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned2_en_5.4.2_3.0_1723923064821.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("catt5_solr_finetunned2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("catt5_solr_finetunned2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|catt5_solr_finetunned2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|918.0 MB| + +## References + +https://huggingface.co/oooriii/catt5-solr-finetunned2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_pipeline_en.md new file mode 100644 index 00000000000000..0bd5e6cdc5bd54 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-catt5_solr_finetunned2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English catt5_solr_finetunned2_pipeline pipeline T5Transformer from oooriii +author: John Snow Labs +name: catt5_solr_finetunned2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`catt5_solr_finetunned2_pipeline` is a English model originally trained by oooriii. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned2_pipeline_en_5.4.2_3.0_1723923137272.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned2_pipeline_en_5.4.2_3.0_1723923137272.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("catt5_solr_finetunned2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("catt5_solr_finetunned2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|catt5_solr_finetunned2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|918.0 MB| + +## References + +https://huggingface.co/oooriii/catt5-solr-finetunned2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-chatcare_sft_en.md b/docs/_posts/ahmedlone127/2024-08-17-chatcare_sft_en.md new file mode 100644 index 00000000000000..10a83d3c468d25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-chatcare_sft_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English chatcare_sft T5Transformer from llllhd +author: John Snow Labs +name: chatcare_sft +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`chatcare_sft` is a English model originally trained by llllhd. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/chatcare_sft_en_5.4.2_3.0_1723931867614.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/chatcare_sft_en_5.4.2_3.0_1723931867614.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("chatcare_sft","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("chatcare_sft", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|chatcare_sft| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/llllhd/ChatCare-SFT \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_en.md b/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_en.md new file mode 100644 index 00000000000000..3eecb1809e04d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English checkpoint_50_epoch_2 T5Transformer from pfactorial +author: John Snow Labs +name: checkpoint_50_epoch_2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_50_epoch_2` is a English model originally trained by pfactorial. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_50_epoch_2_en_5.4.2_3.0_1723920221459.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_50_epoch_2_en_5.4.2_3.0_1723920221459.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("checkpoint_50_epoch_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("checkpoint_50_epoch_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_50_epoch_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pfactorial/checkpoint-50-epoch-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_pipeline_en.md new file mode 100644 index 00000000000000..1bd247d8e624e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-checkpoint_50_epoch_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English checkpoint_50_epoch_2_pipeline pipeline T5Transformer from pfactorial +author: John Snow Labs +name: checkpoint_50_epoch_2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_50_epoch_2_pipeline` is a English model originally trained by pfactorial. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_50_epoch_2_pipeline_en_5.4.2_3.0_1723920272455.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_50_epoch_2_pipeline_en_5.4.2_3.0_1723920272455.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("checkpoint_50_epoch_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("checkpoint_50_epoch_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_50_epoch_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pfactorial/checkpoint-50-epoch-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_en.md b/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_en.md new file mode 100644 index 00000000000000..91287656afbf5f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cnn_summarization_rohan7503 T5Transformer from Rohan7503 +author: John Snow Labs +name: cnn_summarization_rohan7503 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_summarization_rohan7503` is a English model originally trained by Rohan7503. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_summarization_rohan7503_en_5.4.2_3.0_1723933671354.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_summarization_rohan7503_en_5.4.2_3.0_1723933671354.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cnn_summarization_rohan7503","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cnn_summarization_rohan7503", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_summarization_rohan7503| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|326.7 MB| + +## References + +https://huggingface.co/Rohan7503/cnn_summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_pipeline_en.md new file mode 100644 index 00000000000000..fe0f07da506257 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cnn_summarization_rohan7503_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cnn_summarization_rohan7503_pipeline pipeline T5Transformer from Rohan7503 +author: John Snow Labs +name: cnn_summarization_rohan7503_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_summarization_rohan7503_pipeline` is a English model originally trained by Rohan7503. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_summarization_rohan7503_pipeline_en_5.4.2_3.0_1723933693039.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_summarization_rohan7503_pipeline_en_5.4.2_3.0_1723933693039.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cnn_summarization_rohan7503_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cnn_summarization_rohan7503_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_summarization_rohan7503_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|326.7 MB| + +## References + +https://huggingface.co/Rohan7503/cnn_summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_en.md b/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_en.md new file mode 100644 index 00000000000000..c4a1cbeb5e47ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English continue_grammar_model_wmt460 T5Transformer from mika5883 +author: John Snow Labs +name: continue_grammar_model_wmt460 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`continue_grammar_model_wmt460` is a English model originally trained by mika5883. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/continue_grammar_model_wmt460_en_5.4.2_3.0_1723916229889.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/continue_grammar_model_wmt460_en_5.4.2_3.0_1723916229889.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("continue_grammar_model_wmt460","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("continue_grammar_model_wmt460", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|continue_grammar_model_wmt460| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mika5883/continue_grammar_model_wmt460 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_pipeline_en.md new file mode 100644 index 00000000000000..ce56964e1c4214 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-continue_grammar_model_wmt460_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English continue_grammar_model_wmt460_pipeline pipeline T5Transformer from mika5883 +author: John Snow Labs +name: continue_grammar_model_wmt460_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`continue_grammar_model_wmt460_pipeline` is a English model originally trained by mika5883. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/continue_grammar_model_wmt460_pipeline_en_5.4.2_3.0_1723916277952.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/continue_grammar_model_wmt460_pipeline_en_5.4.2_3.0_1723916277952.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("continue_grammar_model_wmt460_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("continue_grammar_model_wmt460_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|continue_grammar_model_wmt460_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mika5883/continue_grammar_model_wmt460 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_en.md new file mode 100644 index 00000000000000..efc4c4887fa1af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_mt5_large_train_instruction0_soapl_v11 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_mt5_large_train_instruction0_soapl_v11 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_mt5_large_train_instruction0_soapl_v11` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_mt5_large_train_instruction0_soapl_v11_en_5.4.2_3.0_1723920384666.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_mt5_large_train_instruction0_soapl_v11_en_5.4.2_3.0_1723920384666.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_mt5_large_train_instruction0_soapl_v11","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_mt5_large_train_instruction0_soapl_v11", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_mt5_large_train_instruction0_soapl_v11| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_mt5-large_train_Instruction0_SOAPL_v11 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline_en.md new file mode 100644 index 00000000000000..d8e88065b7e384 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline_en_5.4.2_3.0_1723920717029.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline_en_5.4.2_3.0_1723920717029.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_mt5_large_train_instruction0_soapl_v11_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_mt5-large_train_Instruction0_SOAPL_v11 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_prompting5_apsol_sup_aug3_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_prompting5_apsol_sup_aug3_en.md new file mode 100644 index 00000000000000..60483422dceb91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_prompting5_apsol_sup_aug3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_sup_aug3 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_sup_aug3 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_sup_aug3` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_aug3_en_5.4.2_3.0_1723924321032.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_aug3_en_5.4.2_3.0_1723924321032.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_sup_aug3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_sup_aug3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_sup_aug3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_SUP_Aug3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_en.md new file mode 100644 index 00000000000000..d5910fce6359bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_pasol_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_pasol_v1_h1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_pasol_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_pasol_v1_h1_en_5.4.2_3.0_1723925553178.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_pasol_v1_h1_en_5.4.2_3.0_1723925553178.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_pasol_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_pasol_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_pasol_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_PASOL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..df34b6878d2f67 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline_en_5.4.2_3.0_1723925734655.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline_en_5.4.2_3.0_1723925734655.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_pasol_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_PASOL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_en.md new file mode 100644 index 00000000000000..b011b2999d7cae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn1_soapl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn1_soapl_v1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn1_soapl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn1_soapl_v1_en_5.4.2_3.0_1723909935455.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn1_soapl_v1_en_5.4.2_3.0_1723909935455.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn1_soapl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn1_soapl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn1_soapl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN1_SOAPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline_en.md new file mode 100644 index 00000000000000..f7a2ea962b5548 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline_en_5.4.2_3.0_1723910109613.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline_en_5.4.2_3.0_1723910109613.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn1_soapl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN1_SOAPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_en.md new file mode 100644 index 00000000000000..49363c143daa21 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn3_soapl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn3_soapl_v1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn3_soapl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn3_soapl_v1_en_5.4.2_3.0_1723926137362.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn3_soapl_v1_en_5.4.2_3.0_1723926137362.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn3_soapl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn3_soapl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn3_soapl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN3_SOAPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline_en.md new file mode 100644 index 00000000000000..4c7c902ba07ea2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline_en_5.4.2_3.0_1723926327368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline_en_5.4.2_3.0_1723926327368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn3_soapl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN3_SOAPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_en.md new file mode 100644 index 00000000000000..b932e995264cab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_spaol_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_spaol_v1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_spaol_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spaol_v1_en_5.4.2_3.0_1723912431800.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spaol_v1_en_5.4.2_3.0_1723912431800.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_spaol_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_spaol_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_spaol_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SPAOL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline_en.md new file mode 100644 index 00000000000000..ec79010a09da65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline_en_5.4.2_3.0_1723912641136.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline_en_5.4.2_3.0_1723912641136.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_spaol_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SPAOL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_en.md new file mode 100644 index 00000000000000..ebd02b80851319 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction4_oaspl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction4_oaspl_v1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction4_oaspl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_oaspl_v1_en_5.4.2_3.0_1723925654592.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_oaspl_v1_en_5.4.2_3.0_1723925654592.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction4_oaspl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction4_oaspl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction4_oaspl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction4_OASPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline_en.md new file mode 100644 index 00000000000000..ff6b457a928cca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline_en_5.4.2_3.0_1723925831811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline_en_5.4.2_3.0_1723925831811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction4_oaspl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction4_OASPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_en.md b/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_en.md new file mode 100644 index 00000000000000..cf038515f37cf9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English derm_t5_small_12 T5Transformer from neal61 +author: John Snow Labs +name: derm_t5_small_12 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`derm_t5_small_12` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/derm_t5_small_12_en_5.4.2_3.0_1723934385760.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/derm_t5_small_12_en_5.4.2_3.0_1723934385760.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("derm_t5_small_12","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("derm_t5_small_12", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|derm_t5_small_12| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.6 MB| + +## References + +https://huggingface.co/neal61/derm-t5-small-12 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_pipeline_en.md new file mode 100644 index 00000000000000..91ed7305d66169 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-derm_t5_small_12_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English derm_t5_small_12_pipeline pipeline T5Transformer from neal61 +author: John Snow Labs +name: derm_t5_small_12_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`derm_t5_small_12_pipeline` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/derm_t5_small_12_pipeline_en_5.4.2_3.0_1723934403024.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/derm_t5_small_12_pipeline_en_5.4.2_3.0_1723934403024.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("derm_t5_small_12_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("derm_t5_small_12_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|derm_t5_small_12_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.6 MB| + +## References + +https://huggingface.co/neal61/derm-t5-small-12 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-det5_base_en.md b/docs/_posts/ahmedlone127/2024-08-17-det5_base_en.md new file mode 100644 index 00000000000000..bc27b6adcbbce7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-det5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English det5_base T5Transformer from helmi0695 +author: John Snow Labs +name: det5_base +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`det5_base` is a English model originally trained by helmi0695. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/det5_base_en_5.4.2_3.0_1723919885488.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/det5_base_en_5.4.2_3.0_1723919885488.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("det5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("det5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|det5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|172.8 MB| + +## References + +https://huggingface.co/helmi0695/det5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-det5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-det5_base_pipeline_en.md new file mode 100644 index 00000000000000..6c71720a33c132 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-det5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English det5_base_pipeline pipeline T5Transformer from helmi0695 +author: John Snow Labs +name: det5_base_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`det5_base_pipeline` is a English model originally trained by helmi0695. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/det5_base_pipeline_en_5.4.2_3.0_1723919945725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/det5_base_pipeline_en_5.4.2_3.0_1723919945725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("det5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("det5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|det5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|172.8 MB| + +## References + +https://huggingface.co/helmi0695/det5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_en.md b/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_en.md new file mode 100644 index 00000000000000..0655122e4b881f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dialot5_base T5Transformer from artemnech +author: John Snow Labs +name: dialot5_base +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dialot5_base` is a English model originally trained by artemnech. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dialot5_base_en_5.4.2_3.0_1723925425850.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dialot5_base_en_5.4.2_3.0_1723925425850.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dialot5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dialot5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dialot5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/artemnech/dialoT5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_pipeline_en.md new file mode 100644 index 00000000000000..71bf474a794aa6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-dialot5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English dialot5_base_pipeline pipeline T5Transformer from artemnech +author: John Snow Labs +name: dialot5_base_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dialot5_base_pipeline` is a English model originally trained by artemnech. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dialot5_base_pipeline_en_5.4.2_3.0_1723925477212.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dialot5_base_pipeline_en_5.4.2_3.0_1723925477212.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("dialot5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("dialot5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dialot5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/artemnech/dialoT5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_en.md b/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_en.md new file mode 100644 index 00000000000000..f935e740d64876 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_0_4_5 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_4_5 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_4_5` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_5_en_5.4.2_3.0_1723912865539.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_5_en_5.4.2_3.0_1723912865539.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_0_4_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_0_4_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_4_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.4-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_pipeline_en.md new file mode 100644 index 00000000000000..bc88fda4f5e338 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-distilled_mt5_small_0_4_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_0_4_5_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_4_5_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_4_5_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_5_pipeline_en_5.4.2_3.0_1723913058818.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_5_pipeline_en_5.4.2_3.0_1723913058818.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_0_4_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_0_4_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_4_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.4-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_en.md b/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_en.md new file mode 100644 index 00000000000000..931626a980bbfb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dpo_anubis_large T5Transformer from danushkhanna +author: John Snow Labs +name: dpo_anubis_large +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dpo_anubis_large` is a English model originally trained by danushkhanna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dpo_anubis_large_en_5.4.2_3.0_1723919752061.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dpo_anubis_large_en_5.4.2_3.0_1723919752061.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dpo_anubis_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dpo_anubis_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dpo_anubis_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.0 MB| + +## References + +https://huggingface.co/danushkhanna/DPO_ANUBIS_LARGE \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_pipeline_en.md new file mode 100644 index 00000000000000..58412d6b150562 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-dpo_anubis_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English dpo_anubis_large_pipeline pipeline T5Transformer from danushkhanna +author: John Snow Labs +name: dpo_anubis_large_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dpo_anubis_large_pipeline` is a English model originally trained by danushkhanna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dpo_anubis_large_pipeline_en_5.4.2_3.0_1723919933281.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dpo_anubis_large_pipeline_en_5.4.2_3.0_1723919933281.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("dpo_anubis_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("dpo_anubis_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dpo_anubis_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.0 MB| + +## References + +https://huggingface.co/danushkhanna/DPO_ANUBIS_LARGE + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-english_mt5_base_5_spider_en.md b/docs/_posts/ahmedlone127/2024-08-17-english_mt5_base_5_spider_en.md new file mode 100644 index 00000000000000..69a3cd45ef0a32 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-english_mt5_base_5_spider_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_mt5_base_5_spider T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_base_5_spider +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_base_5_spider` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_base_5_spider_en_5.4.2_3.0_1723929345345.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_base_5_spider_en_5.4.2_3.0_1723929345345.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_mt5_base_5_spider","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_mt5_base_5_spider", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_base_5_spider| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.1 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-base_5_spider \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-evjvqa_mt5_vit_16_en.md b/docs/_posts/ahmedlone127/2024-08-17-evjvqa_mt5_vit_16_en.md new file mode 100644 index 00000000000000..61f91f79975d46 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-evjvqa_mt5_vit_16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English evjvqa_mt5_vit_16 T5Transformer from luanngo +author: John Snow Labs +name: evjvqa_mt5_vit_16 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`evjvqa_mt5_vit_16` is a English model originally trained by luanngo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/evjvqa_mt5_vit_16_en_5.4.2_3.0_1723931789946.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/evjvqa_mt5_vit_16_en_5.4.2_3.0_1723931789946.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("evjvqa_mt5_vit_16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("evjvqa_mt5_vit_16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|evjvqa_mt5_vit_16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/luanngo/evjvqa_mt5_vit_16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_en.md b/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_en.md new file mode 100644 index 00000000000000..257babf3e6b2f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English experience_extraction_epoc3 T5Transformer from ManulaPankaja +author: John Snow Labs +name: experience_extraction_epoc3 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`experience_extraction_epoc3` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/experience_extraction_epoc3_en_5.4.2_3.0_1723928121340.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/experience_extraction_epoc3_en_5.4.2_3.0_1723928121340.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("experience_extraction_epoc3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("experience_extraction_epoc3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|experience_extraction_epoc3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|984.3 MB| + +## References + +https://huggingface.co/ManulaPankaja/experience_extraction_epoc3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_pipeline_en.md new file mode 100644 index 00000000000000..0693beb8199451 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-experience_extraction_epoc3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English experience_extraction_epoc3_pipeline pipeline T5Transformer from ManulaPankaja +author: John Snow Labs +name: experience_extraction_epoc3_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`experience_extraction_epoc3_pipeline` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/experience_extraction_epoc3_pipeline_en_5.4.2_3.0_1723928178909.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/experience_extraction_epoc3_pipeline_en_5.4.2_3.0_1723928178909.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("experience_extraction_epoc3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("experience_extraction_epoc3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|experience_extraction_epoc3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|984.3 MB| + +## References + +https://huggingface.co/ManulaPankaja/experience_extraction_epoc3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_en.md new file mode 100644 index 00000000000000..54ac3b9d237d83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetuned_t5_small T5Transformer from Palistha +author: John Snow Labs +name: finetuned_t5_small +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_t5_small` is a English model originally trained by Palistha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_t5_small_en_5.4.2_3.0_1723910088978.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_t5_small_en_5.4.2_3.0_1723910088978.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetuned_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetuned_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|341.5 MB| + +## References + +https://huggingface.co/Palistha/finetuned-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..a8d798e4be0690 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-finetuned_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetuned_t5_small_pipeline pipeline T5Transformer from Palistha +author: John Snow Labs +name: finetuned_t5_small_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_t5_small_pipeline` is a English model originally trained by Palistha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_t5_small_pipeline_en_5.4.2_3.0_1723910107697.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_t5_small_pipeline_en_5.4.2_3.0_1723910107697.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetuned_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetuned_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|341.5 MB| + +## References + +https://huggingface.co/Palistha/finetuned-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_en.md new file mode 100644 index 00000000000000..e7cd4571e9281d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_base_text_sql T5Transformer from kunley2 +author: John Snow Labs +name: flan_base_text_sql +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_base_text_sql` is a English model originally trained by kunley2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_base_text_sql_en_5.4.2_3.0_1723925757471.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_base_text_sql_en_5.4.2_3.0_1723925757471.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_base_text_sql","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_base_text_sql", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_base_text_sql| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.6 MB| + +## References + +https://huggingface.co/kunley2/Flan-base-text-sql \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_pipeline_en.md new file mode 100644 index 00000000000000..fabea7818b3c0b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_base_text_sql_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_base_text_sql_pipeline pipeline T5Transformer from kunley2 +author: John Snow Labs +name: flan_base_text_sql_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_base_text_sql_pipeline` is a English model originally trained by kunley2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_base_text_sql_pipeline_en_5.4.2_3.0_1723925940210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_base_text_sql_pipeline_en_5.4.2_3.0_1723925940210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_base_text_sql_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_base_text_sql_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_base_text_sql_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.6 MB| + +## References + +https://huggingface.co/kunley2/Flan-base-text-sql + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_en.md new file mode 100644 index 00000000000000..46a407944ad256 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_answer_generator T5Transformer from currentlyexhausted +author: John Snow Labs +name: flan_t5_answer_generator +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_answer_generator` is a English model originally trained by currentlyexhausted. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_answer_generator_en_5.4.2_3.0_1723938369122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_answer_generator_en_5.4.2_3.0_1723938369122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_answer_generator","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_answer_generator", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_answer_generator| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/currentlyexhausted/flan-t5-answer-generator \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_pipeline_en.md new file mode 100644 index 00000000000000..be963a0aeee46c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_answer_generator_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_answer_generator_pipeline pipeline T5Transformer from currentlyexhausted +author: John Snow Labs +name: flan_t5_answer_generator_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_answer_generator_pipeline` is a English model originally trained by currentlyexhausted. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_answer_generator_pipeline_en_5.4.2_3.0_1723938391663.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_answer_generator_pipeline_en_5.4.2_3.0_1723938391663.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_answer_generator_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_answer_generator_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_answer_generator_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/currentlyexhausted/flan-t5-answer-generator + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_en.md new file mode 100644 index 00000000000000..503fca2c411338 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_analogy_permutation_domain T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_base_analogy_permutation_domain +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_analogy_permutation_domain` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_permutation_domain_en_5.4.2_3.0_1723913511315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_permutation_domain_en_5.4.2_3.0_1723913511315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_analogy_permutation_domain","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_analogy_permutation_domain", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_analogy_permutation_domain| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/flan-t5-base-analogy-permutation-domain \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_pipeline_en.md new file mode 100644 index 00000000000000..88faaba0901597 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_analogy_permutation_domain_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_analogy_permutation_domain_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_base_analogy_permutation_domain_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_analogy_permutation_domain_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_permutation_domain_pipeline_en_5.4.2_3.0_1723913558979.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_permutation_domain_pipeline_en_5.4.2_3.0_1723913558979.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_analogy_permutation_domain_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_analogy_permutation_domain_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_analogy_permutation_domain_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/flan-t5-base-analogy-permutation-domain + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_en.md new file mode 100644 index 00000000000000..3035803ec6fb19 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_summaries_bioarxiv T5Transformer from ViktorDo +author: John Snow Labs +name: flan_t5_base_finetuned_summaries_bioarxiv +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_summaries_bioarxiv` is a English model originally trained by ViktorDo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_summaries_bioarxiv_en_5.4.2_3.0_1723915516371.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_summaries_bioarxiv_en_5.4.2_3.0_1723915516371.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_summaries_bioarxiv","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_summaries_bioarxiv", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_summaries_bioarxiv| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ViktorDo/flan-t5-base-finetuned-summaries-BioArxiv \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_pipeline_en.md new file mode 100644 index 00000000000000..4d6007696babae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_finetuned_summaries_bioarxiv_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_summaries_bioarxiv_pipeline pipeline T5Transformer from ViktorDo +author: John Snow Labs +name: flan_t5_base_finetuned_summaries_bioarxiv_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_summaries_bioarxiv_pipeline` is a English model originally trained by ViktorDo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_summaries_bioarxiv_pipeline_en_5.4.2_3.0_1723915569820.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_summaries_bioarxiv_pipeline_en_5.4.2_3.0_1723915569820.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_summaries_bioarxiv_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_summaries_bioarxiv_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_summaries_bioarxiv_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ViktorDo/flan-t5-base-finetuned-summaries-BioArxiv + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_en.md new file mode 100644 index 00000000000000..bfaa6e5f42f0db --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_gold T5Transformer from prnv13 +author: John Snow Labs +name: flan_t5_base_gold +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_gold` is a English model originally trained by prnv13. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_gold_en_5.4.2_3.0_1723923274676.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_gold_en_5.4.2_3.0_1723923274676.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_gold","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_gold", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_gold| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/prnv13/flan-t5-base-gold \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_pipeline_en.md new file mode 100644 index 00000000000000..bba29a1b1d3922 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_gold_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_gold_pipeline pipeline T5Transformer from prnv13 +author: John Snow Labs +name: flan_t5_base_gold_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_gold_pipeline` is a English model originally trained by prnv13. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_gold_pipeline_en_5.4.2_3.0_1723923324038.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_gold_pipeline_en_5.4.2_3.0_1723923324038.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_gold_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_gold_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_gold_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/prnv13/flan-t5-base-gold + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_en.md new file mode 100644 index 00000000000000..3a84527ea46a7c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_samsum_getrajeev03 T5Transformer from getrajeev03 +author: John Snow Labs +name: flan_t5_base_samsum_getrajeev03 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_samsum_getrajeev03` is a English model originally trained by getrajeev03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_getrajeev03_en_5.4.2_3.0_1723910071201.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_getrajeev03_en_5.4.2_3.0_1723910071201.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_samsum_getrajeev03","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_samsum_getrajeev03", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_samsum_getrajeev03| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/getrajeev03/flan-t5-base-samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_pipeline_en.md new file mode 100644 index 00000000000000..c583f3cc97ce61 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_base_samsum_getrajeev03_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_samsum_getrajeev03_pipeline pipeline T5Transformer from getrajeev03 +author: John Snow Labs +name: flan_t5_base_samsum_getrajeev03_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_samsum_getrajeev03_pipeline` is a English model originally trained by getrajeev03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_getrajeev03_pipeline_en_5.4.2_3.0_1723910119280.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_getrajeev03_pipeline_en_5.4.2_3.0_1723910119280.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_samsum_getrajeev03_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_samsum_getrajeev03_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_samsum_getrajeev03_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/getrajeev03/flan-t5-base-samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_en.md new file mode 100644 index 00000000000000..7cf7492cdaac35 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_dialogsum T5Transformer from Francesco-Acanfora +author: John Snow Labs +name: flan_t5_dialogsum +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_dialogsum` is a English model originally trained by Francesco-Acanfora. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_dialogsum_en_5.4.2_3.0_1723930007680.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_dialogsum_en_5.4.2_3.0_1723930007680.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_dialogsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_dialogsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_dialogsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Francesco-Acanfora/flan-t5-dialogsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_pipeline_en.md new file mode 100644 index 00000000000000..936abd7c6f9a59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_dialogsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_dialogsum_pipeline pipeline T5Transformer from Francesco-Acanfora +author: John Snow Labs +name: flan_t5_dialogsum_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_dialogsum_pipeline` is a English model originally trained by Francesco-Acanfora. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_dialogsum_pipeline_en_5.4.2_3.0_1723930058282.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_dialogsum_pipeline_en_5.4.2_3.0_1723930058282.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_dialogsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_dialogsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_dialogsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Francesco-Acanfora/flan-t5-dialogsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_ep30_lr_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_ep30_lr_en.md new file mode 100644 index 00000000000000..e1011cf41719c6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_ep30_lr_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_medistill_28_ep30_lr T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_28_ep30_lr +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_28_ep30_lr` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_ep30_lr_en_5.4.2_3.0_1723935375486.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_ep30_lr_en_5.4.2_3.0_1723935375486.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_ep30_lr","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_ep30_lr", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_28_ep30_lr| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_28_ep30_lr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_en.md new file mode 100644 index 00000000000000..9fe5c74cfe9dbf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_medistill_28_hyperparameters T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_28_hyperparameters +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_28_hyperparameters` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_en_5.4.2_3.0_1723921711598.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_en_5.4.2_3.0_1723921711598.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_hyperparameters","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_hyperparameters", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_28_hyperparameters| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_28_hyperparameters \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_pipeline_en.md new file mode 100644 index 00000000000000..5011e91fa0ff16 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_medistill_28_hyperparameters_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_medistill_28_hyperparameters_pipeline pipeline T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_28_hyperparameters_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_28_hyperparameters_pipeline` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_pipeline_en_5.4.2_3.0_1723921875592.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_pipeline_en_5.4.2_3.0_1723921875592.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_medistill_28_hyperparameters_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_medistill_28_hyperparameters_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_28_hyperparameters_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_28_hyperparameters + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_svamp_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_svamp_en.md new file mode 100644 index 00000000000000..541d91e2dba2bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_large_svamp_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_svamp T5Transformer from mkhalifa +author: John Snow Labs +name: flan_t5_large_svamp +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_svamp` is a English model originally trained by mkhalifa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_svamp_en_5.4.2_3.0_1723910795509.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_svamp_en_5.4.2_3.0_1723910795509.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_svamp","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_svamp", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_svamp| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/mkhalifa/flan-t5-large-svamp \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_en.md new file mode 100644 index 00000000000000..8b15c467af35dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_retacred_kongo_small T5Transformer from kinshuk-h +author: John Snow Labs +name: flan_t5_retacred_kongo_small +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_retacred_kongo_small` is a English model originally trained by kinshuk-h. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_kongo_small_en_5.4.2_3.0_1723939135068.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_kongo_small_en_5.4.2_3.0_1723939135068.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_retacred_kongo_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_retacred_kongo_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_retacred_kongo_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/kinshuk-h/flan-t5-retacred-kg-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_pipeline_en.md new file mode 100644 index 00000000000000..7874331a4e2910 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_retacred_kongo_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_retacred_kongo_small_pipeline pipeline T5Transformer from kinshuk-h +author: John Snow Labs +name: flan_t5_retacred_kongo_small_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_retacred_kongo_small_pipeline` is a English model originally trained by kinshuk-h. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_kongo_small_pipeline_en_5.4.2_3.0_1723939152296.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_kongo_small_pipeline_en_5.4.2_3.0_1723939152296.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_retacred_kongo_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_retacred_kongo_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_retacred_kongo_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/kinshuk-h/flan-t5-retacred-kg-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_en.md new file mode 100644 index 00000000000000..b5b31fc4a1a8f1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_5_6_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_5_6_cnndm +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_5_6_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_5_6_cnndm_en_5.4.2_3.0_1723927304921.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_5_6_cnndm_en_5.4.2_3.0_1723927304921.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_5_6_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_5_6_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_5_6_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-5-6-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..e208fb75824e43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_5_6_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_5_6_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_5_6_cnndm_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_5_6_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_5_6_cnndm_pipeline_en_5.4.2_3.0_1723927320913.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_5_6_cnndm_pipeline_en_5.4.2_3.0_1723927320913.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_5_6_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_5_6_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_5_6_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-5-6-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_en.md new file mode 100644 index 00000000000000..61af95acaf8fe8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_asap_t4_f2_prompt_adherence T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t4_f2_prompt_adherence +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t4_f2_prompt_adherence` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t4_f2_prompt_adherence_en_5.4.2_3.0_1723916748549.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t4_f2_prompt_adherence_en_5.4.2_3.0_1723916748549.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_asap_t4_f2_prompt_adherence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_asap_t4_f2_prompt_adherence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t4_f2_prompt_adherence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t4_f2_prompt_adherence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_pipeline_en.md new file mode 100644 index 00000000000000..4c2293f02c4cce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_asap_t4_f2_prompt_adherence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_asap_t4_f2_prompt_adherence_pipeline pipeline T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t4_f2_prompt_adherence_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t4_f2_prompt_adherence_pipeline` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t4_f2_prompt_adherence_pipeline_en_5.4.2_3.0_1723916766425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t4_f2_prompt_adherence_pipeline_en_5.4.2_3.0_1723916766425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_asap_t4_f2_prompt_adherence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_asap_t4_f2_prompt_adherence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t4_f2_prompt_adherence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t4_f2_prompt_adherence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_en.md new file mode 100644 index 00000000000000..32fdbb5db2c7d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_4 T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_4 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_4` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_4_en_5.4.2_3.0_1723933654047.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_4_en_5.4.2_3.0_1723933654047.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_pipeline_en.md new file mode 100644 index 00000000000000..8e982424d783b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_finetuned_coqa_v0_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_4_pipeline pipeline T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_4_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_4_pipeline` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_4_pipeline_en_5.4.2_3.0_1723933671098.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_4_pipeline_en_5.4.2_3.0_1723933671098.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_en.md new file mode 100644 index 00000000000000..bbbd146c3433c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_query T5Transformer from gangiswag +author: John Snow Labs +name: flan_t5_small_query +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_query` is a English model originally trained by gangiswag. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_query_en_5.4.2_3.0_1723933244081.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_query_en_5.4.2_3.0_1723933244081.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_query","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_query", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_query| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/gangiswag/flan_t5_small_query \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_pipeline_en.md new file mode 100644 index 00000000000000..a4426a0dc3d419 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_query_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_query_pipeline pipeline T5Transformer from gangiswag +author: John Snow Labs +name: flan_t5_small_query_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_query_pipeline` is a English model originally trained by gangiswag. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_query_pipeline_en_5.4.2_3.0_1723933394156.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_query_pipeline_en_5.4.2_3.0_1723933394156.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_query_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_query_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_query_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/gangiswag/flan_t5_small_query + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_en.md new file mode 100644 index 00000000000000..5bcd349e4ff326 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_samsum_summary T5Transformer from MoralesTP +author: John Snow Labs +name: flan_t5_small_samsum_summary +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_summary` is a English model originally trained by MoralesTP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_summary_en_5.4.2_3.0_1723912240279.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_summary_en_5.4.2_3.0_1723912240279.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_samsum_summary","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_samsum_summary", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_summary| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/MoralesTP/flan-t5-small-samsum-summary \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_pipeline_en.md new file mode 100644 index 00000000000000..26ae3b650ff1a2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_samsum_summary_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_samsum_summary_pipeline pipeline T5Transformer from MoralesTP +author: John Snow Labs +name: flan_t5_small_samsum_summary_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_summary_pipeline` is a English model originally trained by MoralesTP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_summary_pipeline_en_5.4.2_3.0_1723912258389.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_summary_pipeline_en_5.4.2_3.0_1723912258389.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_samsum_summary_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_samsum_summary_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_summary_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/MoralesTP/flan-t5-small-samsum-summary + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_en.md new file mode 100644 index 00000000000000..d6eaf9c27918c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_sat T5Transformer from fiveflow +author: John Snow Labs +name: flan_t5_small_sat +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_sat` is a English model originally trained by fiveflow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_sat_en_5.4.2_3.0_1723932406143.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_sat_en_5.4.2_3.0_1723932406143.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_sat","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_sat", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_sat| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/fiveflow/flan-t5-small-sat \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_pipeline_en.md new file mode 100644 index 00000000000000..91ffb08db2f595 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_small_sat_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_sat_pipeline pipeline T5Transformer from fiveflow +author: John Snow Labs +name: flan_t5_small_sat_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_sat_pipeline` is a English model originally trained by fiveflow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_sat_pipeline_en_5.4.2_3.0_1723932422752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_sat_pipeline_en_5.4.2_3.0_1723932422752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_sat_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_sat_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_sat_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/fiveflow/flan-t5-small-sat + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_en.md new file mode 100644 index 00000000000000..7f21f4ee3a4cf6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_southern_sotho_qg_learningq T5Transformer from sidovic +author: John Snow Labs +name: flan_t5_southern_sotho_qg_learningq +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_southern_sotho_qg_learningq` is a English model originally trained by sidovic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_learningq_en_5.4.2_3.0_1723921141781.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_learningq_en_5.4.2_3.0_1723921141781.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_southern_sotho_qg_learningq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_southern_sotho_qg_learningq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_southern_sotho_qg_learningq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/sidovic/flan-T5-ST-qg-LearningQ \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_pipeline_en.md new file mode 100644 index 00000000000000..9deca503d8f249 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_learningq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_southern_sotho_qg_learningq_pipeline pipeline T5Transformer from sidovic +author: John Snow Labs +name: flan_t5_southern_sotho_qg_learningq_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_southern_sotho_qg_learningq_pipeline` is a English model originally trained by sidovic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_learningq_pipeline_en_5.4.2_3.0_1723921158787.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_learningq_pipeline_en_5.4.2_3.0_1723921158787.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_southern_sotho_qg_learningq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_southern_sotho_qg_learningq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_southern_sotho_qg_learningq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/sidovic/flan-T5-ST-qg-LearningQ + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_en.md new file mode 100644 index 00000000000000..9d2c80cb888ba4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_southern_sotho_qg_squad T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_southern_sotho_qg_squad +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_southern_sotho_qg_squad` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_squad_en_5.4.2_3.0_1723926591947.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_squad_en_5.4.2_3.0_1723926591947.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_southern_sotho_qg_squad","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_southern_sotho_qg_squad", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_southern_sotho_qg_squad| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-T5-ST-qg-SQuAD \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_pipeline_en.md new file mode 100644 index 00000000000000..cff20351d36507 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flan_t5_southern_sotho_qg_squad_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_southern_sotho_qg_squad_pipeline pipeline T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_southern_sotho_qg_squad_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_southern_sotho_qg_squad_pipeline` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_squad_pipeline_en_5.4.2_3.0_1723926641001.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_southern_sotho_qg_squad_pipeline_en_5.4.2_3.0_1723926641001.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_southern_sotho_qg_squad_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_southern_sotho_qg_squad_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_southern_sotho_qg_squad_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-T5-ST-qg-SQuAD + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_en.md b/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_en.md new file mode 100644 index 00000000000000..1da1e14d9af0a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flanb_40000_sp T5Transformer from kyoyanagi +author: John Snow Labs +name: flanb_40000_sp +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flanb_40000_sp` is a English model originally trained by kyoyanagi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flanb_40000_sp_en_5.4.2_3.0_1723932001081.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flanb_40000_sp_en_5.4.2_3.0_1723932001081.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flanb_40000_sp","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flanb_40000_sp", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flanb_40000_sp| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/kyoyanagi/flanb-40000-sp \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_pipeline_en.md new file mode 100644 index 00000000000000..f87f2b2e7270b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flanb_40000_sp_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flanb_40000_sp_pipeline pipeline T5Transformer from kyoyanagi +author: John Snow Labs +name: flanb_40000_sp_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flanb_40000_sp_pipeline` is a English model originally trained by kyoyanagi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flanb_40000_sp_pipeline_en_5.4.2_3.0_1723932054355.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flanb_40000_sp_pipeline_en_5.4.2_3.0_1723932054355.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flanb_40000_sp_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flanb_40000_sp_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flanb_40000_sp_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/kyoyanagi/flanb-40000-sp + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_en.md b/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_en.md new file mode 100644 index 00000000000000..df60bca5994f91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_base_finetuning T5Transformer from tuquyennnn +author: John Snow Labs +name: flant5_base_finetuning +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_base_finetuning` is a English model originally trained by tuquyennnn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_base_finetuning_en_5.4.2_3.0_1723936976192.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_base_finetuning_en_5.4.2_3.0_1723936976192.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_base_finetuning","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_base_finetuning", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_base_finetuning| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tuquyennnn/flant5-base-finetuning \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_pipeline_en.md new file mode 100644 index 00000000000000..ce38c198dbaee1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-flant5_base_finetuning_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_base_finetuning_pipeline pipeline T5Transformer from tuquyennnn +author: John Snow Labs +name: flant5_base_finetuning_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_base_finetuning_pipeline` is a English model originally trained by tuquyennnn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_base_finetuning_pipeline_en_5.4.2_3.0_1723937025847.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_base_finetuning_pipeline_en_5.4.2_3.0_1723937025847.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_base_finetuning_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_base_finetuning_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_base_finetuning_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tuquyennnn/flant5-base-finetuning + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_en.md b/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_en.md new file mode 100644 index 00000000000000..808c8baf9b89e0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English house_ops_t5_small_23 T5Transformer from neal61 +author: John Snow Labs +name: house_ops_t5_small_23 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`house_ops_t5_small_23` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/house_ops_t5_small_23_en_5.4.2_3.0_1723924953954.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/house_ops_t5_small_23_en_5.4.2_3.0_1723924953954.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("house_ops_t5_small_23","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("house_ops_t5_small_23", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|house_ops_t5_small_23| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.1 MB| + +## References + +https://huggingface.co/neal61/house-ops-t5-small-23 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_pipeline_en.md new file mode 100644 index 00000000000000..a2013e38e6f740 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-house_ops_t5_small_23_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English house_ops_t5_small_23_pipeline pipeline T5Transformer from neal61 +author: John Snow Labs +name: house_ops_t5_small_23_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`house_ops_t5_small_23_pipeline` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/house_ops_t5_small_23_pipeline_en_5.4.2_3.0_1723924971998.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/house_ops_t5_small_23_pipeline_en_5.4.2_3.0_1723924971998.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("house_ops_t5_small_23_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("house_ops_t5_small_23_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|house_ops_t5_small_23_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.1 MB| + +## References + +https://huggingface.co/neal61/house-ops-t5-small-23 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_en.md b/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_en.md new file mode 100644 index 00000000000000..055bedf7243509 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English italian_tonga_tonga_islands_english_from_books_model T5Transformer from FiliRezGelly +author: John Snow Labs +name: italian_tonga_tonga_islands_english_from_books_model +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_tonga_tonga_islands_english_from_books_model` is a English model originally trained by FiliRezGelly. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_tonga_tonga_islands_english_from_books_model_en_5.4.2_3.0_1723938109093.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_tonga_tonga_islands_english_from_books_model_en_5.4.2_3.0_1723938109093.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("italian_tonga_tonga_islands_english_from_books_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("italian_tonga_tonga_islands_english_from_books_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_tonga_tonga_islands_english_from_books_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.7 MB| + +## References + +https://huggingface.co/FiliRezGelly/it_to_en_from_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_pipeline_en.md new file mode 100644 index 00000000000000..5bc9697b1e0843 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-italian_tonga_tonga_islands_english_from_books_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English italian_tonga_tonga_islands_english_from_books_model_pipeline pipeline T5Transformer from FiliRezGelly +author: John Snow Labs +name: italian_tonga_tonga_islands_english_from_books_model_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_tonga_tonga_islands_english_from_books_model_pipeline` is a English model originally trained by FiliRezGelly. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_tonga_tonga_islands_english_from_books_model_pipeline_en_5.4.2_3.0_1723938128657.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_tonga_tonga_islands_english_from_books_model_pipeline_en_5.4.2_3.0_1723938128657.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("italian_tonga_tonga_islands_english_from_books_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("italian_tonga_tonga_islands_english_from_books_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_tonga_tonga_islands_english_from_books_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.7 MB| + +## References + +https://huggingface.co/FiliRezGelly/it_to_en_from_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_en.md new file mode 100644 index 00000000000000..0316f166cd62dd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ke_t5_small_nikl_summarization T5Transformer from chunwoolee0 +author: John Snow Labs +name: ke_t5_small_nikl_summarization +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_small_nikl_summarization` is a English model originally trained by chunwoolee0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_small_nikl_summarization_en_5.4.2_3.0_1723913936200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_small_nikl_summarization_en_5.4.2_3.0_1723913936200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ke_t5_small_nikl_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ke_t5_small_nikl_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_small_nikl_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|273.6 MB| + +## References + +https://huggingface.co/chunwoolee0/ke_t5_small_nikl_summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_pipeline_en.md new file mode 100644 index 00000000000000..bfd497968ed79e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-ke_t5_small_nikl_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ke_t5_small_nikl_summarization_pipeline pipeline T5Transformer from chunwoolee0 +author: John Snow Labs +name: ke_t5_small_nikl_summarization_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_small_nikl_summarization_pipeline` is a English model originally trained by chunwoolee0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_small_nikl_summarization_pipeline_en_5.4.2_3.0_1723914029654.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_small_nikl_summarization_pipeline_en_5.4.2_3.0_1723914029654.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ke_t5_small_nikl_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ke_t5_small_nikl_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_small_nikl_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|273.6 MB| + +## References + +https://huggingface.co/chunwoolee0/ke_t5_small_nikl_summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_en.md b/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_en.md new file mode 100644 index 00000000000000..13e602fd3f15b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kgt5v2_base_wikikg90mv2 T5Transformer from apoorvumang +author: John Snow Labs +name: kgt5v2_base_wikikg90mv2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kgt5v2_base_wikikg90mv2` is a English model originally trained by apoorvumang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kgt5v2_base_wikikg90mv2_en_5.4.2_3.0_1723922545554.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kgt5v2_base_wikikg90mv2_en_5.4.2_3.0_1723922545554.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kgt5v2_base_wikikg90mv2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kgt5v2_base_wikikg90mv2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kgt5v2_base_wikikg90mv2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/apoorvumang/kgt5v2-base-wikikg90mv2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_pipeline_en.md new file mode 100644 index 00000000000000..178943fcbf513b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kgt5v2_base_wikikg90mv2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kgt5v2_base_wikikg90mv2_pipeline pipeline T5Transformer from apoorvumang +author: John Snow Labs +name: kgt5v2_base_wikikg90mv2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kgt5v2_base_wikikg90mv2_pipeline` is a English model originally trained by apoorvumang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kgt5v2_base_wikikg90mv2_pipeline_en_5.4.2_3.0_1723922597838.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kgt5v2_base_wikikg90mv2_pipeline_en_5.4.2_3.0_1723922597838.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kgt5v2_base_wikikg90mv2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kgt5v2_base_wikikg90mv2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kgt5v2_base_wikikg90mv2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/apoorvumang/kgt5v2-base-wikikg90mv2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_en.md new file mode 100644 index 00000000000000..cd6ba57a5ad785 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_osapl T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_osapl +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_osapl` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_osapl_en_5.4.2_3.0_1723926635353.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_osapl_en_5.4.2_3.0_1723926635353.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_osapl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_osapl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_osapl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OSAPL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_pipeline_en.md new file mode 100644 index 00000000000000..a889b2fcbb1286 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_osapl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_osapl_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_osapl_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_osapl_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_osapl_pipeline_en_5.4.2_3.0_1723926814912.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_osapl_pipeline_en_5.4.2_3.0_1723926814912.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_osapl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_osapl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_osapl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OSAPL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_en.md new file mode 100644 index 00000000000000..0ab6ef1f856a83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_spaol T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_spaol +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_spaol` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spaol_en_5.4.2_3.0_1723921897005.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spaol_en_5.4.2_3.0_1723921897005.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_spaol","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_spaol", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_spaol| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SPAOL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_pipeline_en.md new file mode 100644 index 00000000000000..9e387266ad835e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_spaol_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_spaol_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_spaol_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_spaol_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spaol_pipeline_en_5.4.2_3.0_1723922096168.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spaol_pipeline_en_5.4.2_3.0_1723922096168.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_spaol_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_spaol_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_spaol_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SPAOL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_en.md new file mode 100644 index 00000000000000..41e53fab2c2138 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_posal_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_posal_v5 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_posal_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v5_en_5.4.2_3.0_1723919179216.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v5_en_5.4.2_3.0_1723919179216.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_posal_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POSAL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_pipeline_en.md new file mode 100644 index 00000000000000..68d6c0472d1ca2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_posal_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_posal_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_posal_v5_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_posal_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v5_pipeline_en_5.4.2_3.0_1723919350087.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v5_pipeline_en_5.4.2_3.0_1723919350087.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_posal_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_posal_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_posal_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POSAL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_en.md new file mode 100644 index 00000000000000..dc8b4f5169b8f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_v3 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_v3 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_v3` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v3_en_5.4.2_3.0_1723912430923.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v3_en_5.4.2_3.0_1723912430923.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_pipeline_en.md new file mode 100644 index 00000000000000..165ab8fae6e9fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_sapol_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_v3_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_v3_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_v3_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v3_pipeline_en_5.4.2_3.0_1723912634049.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v3_pipeline_en_5.4.2_3.0_1723912634049.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_sapol_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_sapol_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_spaol_v2_en.md b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_spaol_v2_en.md new file mode 100644 index 00000000000000..315ad56af120d4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-kltn_coqe_vit5_total_spaol_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_spaol_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_spaol_v2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_spaol_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v2_en_5.4.2_3.0_1723931270463.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v2_en_5.4.2_3.0_1723931270463.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_spaol_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_spaol_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_spaol_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SPAOL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_en.md b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_en.md new file mode 100644 index 00000000000000..f4a25451d420cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_multitask_french_german T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_french_german +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_french_german` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_french_german_en_5.4.2_3.0_1723921876720.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_french_german_en_5.4.2_3.0_1723921876720.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_multitask_french_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_multitask_french_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_french_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_fr_de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_pipeline_en.md new file mode 100644 index 00000000000000..a62c3182dfa4ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_multitask_french_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_multitask_french_german_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_french_german_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_french_german_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_french_german_pipeline_en_5.4.2_3.0_1723921939814.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_french_german_pipeline_en_5.4.2_3.0_1723921939814.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_multitask_french_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_multitask_french_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_french_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_fr_de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_en.md b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_en.md new file mode 100644 index 00000000000000..8007861c90ed91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_summ_multitask_english T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_summ_multitask_english +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_summ_multitask_english` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_summ_multitask_english_en_5.4.2_3.0_1723913158274.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_summ_multitask_english_en_5.4.2_3.0_1723913158274.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_summ_multitask_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_summ_multitask_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_summ_multitask_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.7 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_summ_multitask_en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_pipeline_en.md new file mode 100644 index 00000000000000..e4ffdcd5b8663b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-legal_t5_small_summ_multitask_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_summ_multitask_english_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_summ_multitask_english_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_summ_multitask_english_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_summ_multitask_english_pipeline_en_5.4.2_3.0_1723913220617.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_summ_multitask_english_pipeline_en_5.4.2_3.0_1723913220617.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_summ_multitask_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_summ_multitask_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_summ_multitask_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.7 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_summ_multitask_en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_en.md b/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_en.md new file mode 100644 index 00000000000000..3376242d974bbb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lever_wikitq_codex T5Transformer from niansong1996 +author: John Snow Labs +name: lever_wikitq_codex +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lever_wikitq_codex` is a English model originally trained by niansong1996. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lever_wikitq_codex_en_5.4.2_3.0_1723928831637.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lever_wikitq_codex_en_5.4.2_3.0_1723928831637.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lever_wikitq_codex","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lever_wikitq_codex", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lever_wikitq_codex| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/niansong1996/lever-wikitq-codex \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_pipeline_en.md new file mode 100644 index 00000000000000..6f5f13a1c17efe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-lever_wikitq_codex_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lever_wikitq_codex_pipeline pipeline T5Transformer from niansong1996 +author: John Snow Labs +name: lever_wikitq_codex_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lever_wikitq_codex_pipeline` is a English model originally trained by niansong1996. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lever_wikitq_codex_pipeline_en_5.4.2_3.0_1723928983127.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lever_wikitq_codex_pipeline_en_5.4.2_3.0_1723928983127.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lever_wikitq_codex_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lever_wikitq_codex_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lever_wikitq_codex_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/niansong1996/lever-wikitq-codex + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_en.md b/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_en.md new file mode 100644 index 00000000000000..c32014f9935e43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English long_t5lephone_5000 T5Transformer from Splend1dchan +author: John Snow Labs +name: long_t5lephone_5000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5lephone_5000` is a English model originally trained by Splend1dchan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5lephone_5000_en_5.4.2_3.0_1723909714912.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5lephone_5000_en_5.4.2_3.0_1723909714912.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("long_t5lephone_5000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("long_t5lephone_5000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5lephone_5000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Splend1dchan/long-t5lephone-5000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_pipeline_en.md new file mode 100644 index 00000000000000..dbc224f04e9f18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-long_t5lephone_5000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English long_t5lephone_5000_pipeline pipeline T5Transformer from Splend1dchan +author: John Snow Labs +name: long_t5lephone_5000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5lephone_5000_pipeline` is a English model originally trained by Splend1dchan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5lephone_5000_pipeline_en_5.4.2_3.0_1723909761418.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5lephone_5000_pipeline_en_5.4.2_3.0_1723909761418.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("long_t5lephone_5000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("long_t5lephone_5000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5lephone_5000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Splend1dchan/long-t5lephone-5000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-low8_en.md b/docs/_posts/ahmedlone127/2024-08-17-low8_en.md new file mode 100644 index 00000000000000..1e60d57a0dc444 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-low8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English low8 T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: low8 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`low8` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/low8_en_5.4.2_3.0_1723937493060.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/low8_en_5.4.2_3.0_1723937493060.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("low8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("low8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|low8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|986.3 MB| + +## References + +https://huggingface.co/AliGhiasvand86/low8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-low8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-low8_pipeline_en.md new file mode 100644 index 00000000000000..38904fdee95e98 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-low8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English low8_pipeline pipeline T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: low8_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`low8_pipeline` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/low8_pipeline_en_5.4.2_3.0_1723937549155.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/low8_pipeline_en_5.4.2_3.0_1723937549155.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("low8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("low8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|low8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|986.3 MB| + +## References + +https://huggingface.co/AliGhiasvand86/low8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_en.md b/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_en.md new file mode 100644 index 00000000000000..120688f1f97b97 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_base_boun_first_2012 T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_first_2012 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_first_2012` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_en_5.4.2_3.0_1723932657294.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_en_5.4.2_3.0_1723932657294.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_base_boun_first_2012","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_base_boun_first_2012", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_first_2012| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_first_2012 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_pipeline_en.md new file mode 100644 index 00000000000000..7dbdfc0fab84ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-md_mt5_base_boun_first_2012_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_base_boun_first_2012_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_first_2012_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_first_2012_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_pipeline_en_5.4.2_3.0_1723932830774.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_pipeline_en_5.4.2_3.0_1723932830774.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_base_boun_first_2012_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_base_boun_first_2012_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_first_2012_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_first_2012 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_en.md new file mode 100644 index 00000000000000..da15e708694aa2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_frquad_ae_trimmed_50000 T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_frquad_ae_trimmed_50000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_frquad_ae_trimmed_50000` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_frquad_ae_trimmed_50000_en_5.4.2_3.0_1723931893957.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_frquad_ae_trimmed_50000_en_5.4.2_3.0_1723931893957.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_frquad_ae_trimmed_50000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_frquad_ae_trimmed_50000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_frquad_ae_trimmed_50000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-frquad-ae-trimmed-50000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_pipeline_en.md new file mode 100644 index 00000000000000..47a5d88c369cfc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_base_frquad_ae_trimmed_50000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_frquad_ae_trimmed_50000_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_frquad_ae_trimmed_50000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_frquad_ae_trimmed_50000_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_frquad_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1723931957814.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_frquad_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1723931957814.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_frquad_ae_trimmed_50000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_frquad_ae_trimmed_50000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_frquad_ae_trimmed_50000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-frquad-ae-trimmed-50000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_en.md new file mode 100644 index 00000000000000..09f2aec7fa2054 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_deed_sum_hasanur525 T5Transformer from Hasanur525 +author: John Snow Labs +name: mt5_deed_sum_hasanur525 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_deed_sum_hasanur525` is a English model originally trained by Hasanur525. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_deed_sum_hasanur525_en_5.4.2_3.0_1723913999874.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_deed_sum_hasanur525_en_5.4.2_3.0_1723913999874.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_deed_sum_hasanur525","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_deed_sum_hasanur525", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_deed_sum_hasanur525| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/Hasanur525/mt5_deed_sum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_pipeline_en.md new file mode 100644 index 00000000000000..d05bdcd7200b19 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_deed_sum_hasanur525_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_deed_sum_hasanur525_pipeline pipeline T5Transformer from Hasanur525 +author: John Snow Labs +name: mt5_deed_sum_hasanur525_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_deed_sum_hasanur525_pipeline` is a English model originally trained by Hasanur525. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_deed_sum_hasanur525_pipeline_en_5.4.2_3.0_1723914214993.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_deed_sum_hasanur525_pipeline_en_5.4.2_3.0_1723914214993.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_deed_sum_hasanur525_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_deed_sum_hasanur525_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_deed_sum_hasanur525_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/Hasanur525/mt5_deed_sum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_de.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_de.md new file mode 100644 index 00000000000000..43603cbf1cf564 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_de.md @@ -0,0 +1,86 @@ +--- +layout: model +title: German mt5_simple_german_corpus T5Transformer from DEplain +author: John Snow Labs +name: mt5_simple_german_corpus +date: 2024-08-17 +tags: [de, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_simple_german_corpus` is a German model originally trained by DEplain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_simple_german_corpus_de_5.4.2_3.0_1723918624385.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_simple_german_corpus_de_5.4.2_3.0_1723918624385.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_simple_german_corpus","de") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_simple_german_corpus", "de") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_simple_german_corpus| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|de| +|Size:|2.3 GB| + +## References + +https://huggingface.co/DEplain/mt5-simple-german-corpus \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_pipeline_de.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_pipeline_de.md new file mode 100644 index 00000000000000..a88c6999033573 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_simple_german_corpus_pipeline_de.md @@ -0,0 +1,69 @@ +--- +layout: model +title: German mt5_simple_german_corpus_pipeline pipeline T5Transformer from DEplain +author: John Snow Labs +name: mt5_simple_german_corpus_pipeline +date: 2024-08-17 +tags: [de, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_simple_german_corpus_pipeline` is a German model originally trained by DEplain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_simple_german_corpus_pipeline_de_5.4.2_3.0_1723918784036.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_simple_german_corpus_pipeline_de_5.4.2_3.0_1723918784036.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_simple_german_corpus_pipeline", lang = "de") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_simple_german_corpus_pipeline", lang = "de") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_simple_german_corpus_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|de| +|Size:|2.3 GB| + +## References + +https://huggingface.co/DEplain/mt5-simple-german-corpus + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_en.md new file mode 100644 index 00000000000000..bacc2c24ff896d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_konic T5Transformer from Konic +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_konic +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_konic` is a English model originally trained by Konic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_konic_en_5.4.2_3.0_1723922049035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_konic_en_5.4.2_3.0_1723922049035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_konic","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_konic", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_konic| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Konic/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_pipeline_en.md new file mode 100644 index 00000000000000..6fade79dba1c51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_konic_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_konic_pipeline pipeline T5Transformer from Konic +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_konic_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_konic_pipeline` is a English model originally trained by Konic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_konic_pipeline_en_5.4.2_3.0_1723922181995.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_konic_pipeline_en_5.4.2_3.0_1723922181995.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_konic_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_konic_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_konic_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Konic/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_en.md new file mode 100644 index 00000000000000..a7d83a0d8e4d92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_pnr_svc T5Transformer from pnr-svc +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_pnr_svc +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_pnr_svc` is a English model originally trained by pnr-svc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_pnr_svc_en_5.4.2_3.0_1723935053026.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_pnr_svc_en_5.4.2_3.0_1723935053026.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_pnr_svc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_pnr_svc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_pnr_svc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/pnr-svc/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline_en.md new file mode 100644 index 00000000000000..f6efa73f2fe596 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline pipeline T5Transformer from pnr-svc +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline` is a English model originally trained by pnr-svc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline_en_5.4.2_3.0_1723935173461.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline_en_5.4.2_3.0_1723935173461.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_pnr_svc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/pnr-svc/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_en.md new file mode 100644 index 00000000000000..0991f7e72c3de4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_radhikabansal T5Transformer from radhikabansal +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_radhikabansal +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_radhikabansal` is a English model originally trained by radhikabansal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_radhikabansal_en_5.4.2_3.0_1723935327483.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_radhikabansal_en_5.4.2_3.0_1723935327483.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_radhikabansal","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_radhikabansal", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_radhikabansal| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/radhikabansal/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline_en.md new file mode 100644 index 00000000000000..10adec18713b1e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline pipeline T5Transformer from radhikabansal +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline` is a English model originally trained by radhikabansal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline_en_5.4.2_3.0_1723935439734.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline_en_5.4.2_3.0_1723935439734.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_radhikabansal_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/radhikabansal/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_en.md new file mode 100644 index 00000000000000..ac71934eaa41d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_sasuke T5Transformer from sasuke +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_sasuke +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_sasuke` is a English model originally trained by sasuke. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_sasuke_en_5.4.2_3.0_1723918589115.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_sasuke_en_5.4.2_3.0_1723918589115.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_sasuke","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_sasuke", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_sasuke| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/sasuke/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline_en.md new file mode 100644 index 00000000000000..c8541066c37e8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline pipeline T5Transformer from sasuke +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline` is a English model originally trained by sasuke. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline_en_5.4.2_3.0_1723918706631.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline_en_5.4.2_3.0_1723918706631.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_sasuke_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/sasuke/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_en.md new file mode 100644 index 00000000000000..d10e92c846ea61 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_thetang T5Transformer from Thetang +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_thetang +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_thetang` is a English model originally trained by Thetang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_thetang_en_5.4.2_3.0_1723918707884.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_thetang_en_5.4.2_3.0_1723918707884.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_thetang","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_thetang", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_thetang| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Thetang/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_pipeline_en.md new file mode 100644 index 00000000000000..9446e422325c9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_finetuned_amazon_english_spanish_thetang_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_thetang_pipeline pipeline T5Transformer from Thetang +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_thetang_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_thetang_pipeline` is a English model originally trained by Thetang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_thetang_pipeline_en_5.4.2_3.0_1723918812407.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_thetang_pipeline_en_5.4.2_3.0_1723918812407.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_thetang_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_thetang_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_thetang_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Thetang/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_en.md new file mode 100644 index 00000000000000..2a147dbd1fb6c3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_irish_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_irish_10k +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_irish_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_irish_10k_en_5.4.2_3.0_1723937000781.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_irish_10k_en_5.4.2_3.0_1723937000781.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_irish_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_irish_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_irish_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ga-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_pipeline_en.md new file mode 100644 index 00000000000000..a0e0f104b56641 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_irish_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_irish_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_irish_10k_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_irish_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_irish_10k_pipeline_en_5.4.2_3.0_1723937179398.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_irish_10k_pipeline_en_5.4.2_3.0_1723937179398.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_irish_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_irish_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_irish_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ga-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_en.md new file mode 100644 index 00000000000000..6e2a99bf03a406 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_120000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_120000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_120000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_120000_en_5.4.2_3.0_1723932607735.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_120000_en_5.4.2_3.0_1723932607735.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_120000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_120000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_120000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|714.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-120000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_pipeline_en.md new file mode 100644 index 00000000000000..27dad9b46aca9b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_120000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_120000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_120000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_120000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_120000_pipeline_en_5.4.2_3.0_1723932655027.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_120000_pipeline_en_5.4.2_3.0_1723932655027.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_120000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_120000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_120000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|714.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-120000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_en.md new file mode 100644 index 00000000000000..afa94bd4ac9f60 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_15000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_15000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_15000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_15000_en_5.4.2_3.0_1723919849645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_15000_en_5.4.2_3.0_1723919849645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_15000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_15000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_15000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|252.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-15000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_pipeline_en.md new file mode 100644 index 00000000000000..5fee3b70d7f5cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_ruquad_qa_trimmed_russian_15000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_15000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_15000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_15000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_15000_pipeline_en_5.4.2_3.0_1723919862072.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_15000_pipeline_en_5.4.2_3.0_1723919862072.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_15000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_15000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_15000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|252.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-15000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_en.md new file mode 100644 index 00000000000000..5ae94c3c437b04 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_serbian_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_serbian_10k +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_serbian_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_serbian_10k_en_5.4.2_3.0_1723923445055.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_serbian_10k_en_5.4.2_3.0_1723923445055.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_serbian_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_serbian_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_serbian_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-sr-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_pipeline_en.md new file mode 100644 index 00000000000000..245140a6d9ac85 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_serbian_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_serbian_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_serbian_10k_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_serbian_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_serbian_10k_pipeline_en_5.4.2_3.0_1723923605131.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_serbian_10k_pipeline_en_5.4.2_3.0_1723923605131.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_serbian_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_serbian_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_serbian_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-sr-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_en.md new file mode 100644 index 00000000000000..115b75a5772b24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_english_60000_squad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_60000_squad_qg +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_60000_squad_qg` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qg_en_5.4.2_3.0_1723915986531.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qg_en_5.4.2_3.0_1723915986531.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_english_60000_squad_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_english_60000_squad_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_60000_squad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|481.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-60000-squad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_pipeline_en.md new file mode 100644 index 00000000000000..9b9652e9b98353 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_english_60000_squad_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_english_60000_squad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_60000_squad_qg_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_60000_squad_qg_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qg_pipeline_en_5.4.2_3.0_1723916012433.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qg_pipeline_en_5.4.2_3.0_1723916012433.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_english_60000_squad_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_english_60000_squad_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_60000_squad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|481.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-60000-squad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_it.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_it.md new file mode 100644 index 00000000000000..a104310aeaaa9b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_it.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_itquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_itquad_qg +date: 2024-08-17 +tags: [it, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_itquad_qg` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_itquad_qg_it_5.4.2_3.0_1723913566630.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_itquad_qg_it_5.4.2_3.0_1723913566630.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_itquad_qg","it") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_itquad_qg", "it") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_itquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|it| +|Size:|701.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-itquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_pipeline_it.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_pipeline_it.md new file mode 100644 index 00000000000000..a86cc93bb6d04c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_italian_itquad_qg_pipeline_it.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_itquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_itquad_qg_pipeline +date: 2024-08-17 +tags: [it, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_itquad_qg_pipeline` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_itquad_qg_pipeline_it_5.4.2_3.0_1723913611509.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_itquad_qg_pipeline_it_5.4.2_3.0_1723913611509.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_itquad_qg_pipeline", lang = "it") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_itquad_qg_pipeline", lang = "it") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_itquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|it| +|Size:|701.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-itquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_en.md new file mode 100644 index 00000000000000..e71635e61619ad --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_korean_15000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_15000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_15000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_15000_en_5.4.2_3.0_1723932510375.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_15000_en_5.4.2_3.0_1723932510375.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_15000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_15000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_15000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|130.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-15000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_pipeline_en.md new file mode 100644 index 00000000000000..5164d7c5d112d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_korean_15000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_korean_15000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_15000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_15000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_15000_pipeline_en_5.4.2_3.0_1723932555832.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_15000_pipeline_en_5.4.2_3.0_1723932555832.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_korean_15000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_korean_15000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_15000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|130.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-15000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_en.md new file mode 100644 index 00000000000000..c2bd24d3a08156 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_russian_60000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_60000 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_60000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_60000_en_5.4.2_3.0_1723933768924.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_60000_en_5.4.2_3.0_1723933768924.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_60000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_60000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_60000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|262.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-60000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_pipeline_en.md new file mode 100644 index 00000000000000..3860674f049105 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_small_trimmed_russian_60000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_russian_60000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_60000_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_60000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_60000_pipeline_en_5.4.2_3.0_1723933859752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_60000_pipeline_en_5.4.2_3.0_1723933859752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_russian_60000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_russian_60000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_60000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|262.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-60000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_en.md new file mode 100644 index 00000000000000..873da4d281ca3e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_thaisum_model T5Transformer from Mayypeeya +author: John Snow Labs +name: mt5_thaisum_model +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thaisum_model` is a English model originally trained by Mayypeeya. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thaisum_model_en_5.4.2_3.0_1723933775693.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thaisum_model_en_5.4.2_3.0_1723933775693.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_thaisum_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_thaisum_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thaisum_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Mayypeeya/mt5_thaisum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_pipeline_en.md new file mode 100644 index 00000000000000..b5ab1f8ae09f91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_thaisum_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_thaisum_model_pipeline pipeline T5Transformer from Mayypeeya +author: John Snow Labs +name: mt5_thaisum_model_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thaisum_model_pipeline` is a English model originally trained by Mayypeeya. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thaisum_model_pipeline_en_5.4.2_3.0_1723933940403.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thaisum_model_pipeline_en_5.4.2_3.0_1723933940403.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_thaisum_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_thaisum_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thaisum_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Mayypeeya/mt5_thaisum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_en.md new file mode 100644 index 00000000000000..530ca50cbe0719 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_xl_nestoralvaro_normail T5Transformer from bg79-v23-bidata-ntnu +author: John Snow Labs +name: mt5_xl_nestoralvaro_normail +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_xl_nestoralvaro_normail` is a English model originally trained by bg79-v23-bidata-ntnu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_xl_nestoralvaro_normail_en_5.4.2_3.0_1723938152835.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_xl_nestoralvaro_normail_en_5.4.2_3.0_1723938152835.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_xl_nestoralvaro_normail","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_xl_nestoralvaro_normail", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_xl_nestoralvaro_normail| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|6.0 MB| + +## References + +https://huggingface.co/bg79-v23-bidata-ntnu/mt5_xl-nestoralvaro-normail \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_pipeline_en.md new file mode 100644 index 00000000000000..7819d127743eb5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mt5_xl_nestoralvaro_normail_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_xl_nestoralvaro_normail_pipeline pipeline T5Transformer from bg79-v23-bidata-ntnu +author: John Snow Labs +name: mt5_xl_nestoralvaro_normail_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_xl_nestoralvaro_normail_pipeline` is a English model originally trained by bg79-v23-bidata-ntnu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_xl_nestoralvaro_normail_pipeline_en_5.4.2_3.0_1723938162854.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_xl_nestoralvaro_normail_pipeline_en_5.4.2_3.0_1723938162854.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_xl_nestoralvaro_normail_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_xl_nestoralvaro_normail_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_xl_nestoralvaro_normail_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|6.0 MB| + +## References + +https://huggingface.co/bg79-v23-bidata-ntnu/mt5_xl-nestoralvaro-normail + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-mulitplydivision_en.md b/docs/_posts/ahmedlone127/2024-08-17-mulitplydivision_en.md new file mode 100644 index 00000000000000..ce9f074a59c2ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-mulitplydivision_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mulitplydivision T5Transformer from sammanamgain +author: John Snow Labs +name: mulitplydivision +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mulitplydivision` is a English model originally trained by sammanamgain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mulitplydivision_en_5.4.2_3.0_1723918093944.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mulitplydivision_en_5.4.2_3.0_1723918093944.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mulitplydivision","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mulitplydivision", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mulitplydivision| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/sammanamgain/Mulitplydivision \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_en.md b/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_en.md new file mode 100644 index 00000000000000..8d2ee1e080e294 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English munna_bhai_mbbs_model_final T5Transformer from sharifMunna +author: John Snow Labs +name: munna_bhai_mbbs_model_final +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`munna_bhai_mbbs_model_final` is a English model originally trained by sharifMunna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/munna_bhai_mbbs_model_final_en_5.4.2_3.0_1723922194776.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/munna_bhai_mbbs_model_final_en_5.4.2_3.0_1723922194776.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("munna_bhai_mbbs_model_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("munna_bhai_mbbs_model_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|munna_bhai_mbbs_model_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|940.9 MB| + +## References + +https://huggingface.co/sharifMunna/munna_bhai_mbbs_model_final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_pipeline_en.md new file mode 100644 index 00000000000000..40c6c9b09c005e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-munna_bhai_mbbs_model_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English munna_bhai_mbbs_model_final_pipeline pipeline T5Transformer from sharifMunna +author: John Snow Labs +name: munna_bhai_mbbs_model_final_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`munna_bhai_mbbs_model_final_pipeline` is a English model originally trained by sharifMunna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/munna_bhai_mbbs_model_final_pipeline_en_5.4.2_3.0_1723922259604.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/munna_bhai_mbbs_model_final_pipeline_en_5.4.2_3.0_1723922259604.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("munna_bhai_mbbs_model_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("munna_bhai_mbbs_model_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|munna_bhai_mbbs_model_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|940.9 MB| + +## References + +https://huggingface.co/sharifMunna/munna_bhai_mbbs_model_final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_en.md b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_en.md new file mode 100644 index 00000000000000..1a2e3e94940d06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nep_spell_mt5_small_01 T5Transformer from duraad +author: John Snow Labs +name: nep_spell_mt5_small_01 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nep_spell_mt5_small_01` is a English model originally trained by duraad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_01_en_5.4.2_3.0_1723924911594.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_01_en_5.4.2_3.0_1723924911594.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nep_spell_mt5_small_01","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nep_spell_mt5_small_01", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nep_spell_mt5_small_01| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/duraad/nep-spell-mt5-small-01 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_pipeline_en.md new file mode 100644 index 00000000000000..fa11fbc53739ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_01_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nep_spell_mt5_small_01_pipeline pipeline T5Transformer from duraad +author: John Snow Labs +name: nep_spell_mt5_small_01_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nep_spell_mt5_small_01_pipeline` is a English model originally trained by duraad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_01_pipeline_en_5.4.2_3.0_1723925197895.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_01_pipeline_en_5.4.2_3.0_1723925197895.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nep_spell_mt5_small_01_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nep_spell_mt5_small_01_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nep_spell_mt5_small_01_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/duraad/nep-spell-mt5-small-01 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_en.md b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_en.md new file mode 100644 index 00000000000000..f253f8afe88aa5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nep_spell_mt5_small_1 T5Transformer from duraad +author: John Snow Labs +name: nep_spell_mt5_small_1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nep_spell_mt5_small_1` is a English model originally trained by duraad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_1_en_5.4.2_3.0_1723910761934.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_1_en_5.4.2_3.0_1723910761934.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nep_spell_mt5_small_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nep_spell_mt5_small_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nep_spell_mt5_small_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/duraad/nep-spell-mt5-small-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_pipeline_en.md new file mode 100644 index 00000000000000..a760f9f79abc86 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-nep_spell_mt5_small_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nep_spell_mt5_small_1_pipeline pipeline T5Transformer from duraad +author: John Snow Labs +name: nep_spell_mt5_small_1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nep_spell_mt5_small_1_pipeline` is a English model originally trained by duraad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_1_pipeline_en_5.4.2_3.0_1723911044726.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nep_spell_mt5_small_1_pipeline_en_5.4.2_3.0_1723911044726.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nep_spell_mt5_small_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nep_spell_mt5_small_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nep_spell_mt5_small_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/duraad/nep-spell-mt5-small-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_en.md b/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_en.md new file mode 100644 index 00000000000000..ce764272468ebe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English news_tonga_tonga_islands_speech_t5_summarizer T5Transformer from wrongserenity +author: John Snow Labs +name: news_tonga_tonga_islands_speech_t5_summarizer +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`news_tonga_tonga_islands_speech_t5_summarizer` is a English model originally trained by wrongserenity. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/news_tonga_tonga_islands_speech_t5_summarizer_en_5.4.2_3.0_1723916556273.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/news_tonga_tonga_islands_speech_t5_summarizer_en_5.4.2_3.0_1723916556273.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("news_tonga_tonga_islands_speech_t5_summarizer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("news_tonga_tonga_islands_speech_t5_summarizer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|news_tonga_tonga_islands_speech_t5_summarizer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|976.8 MB| + +## References + +https://huggingface.co/wrongserenity/news_to_speech_t5_summarizer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_pipeline_en.md new file mode 100644 index 00000000000000..e85b06d468a66c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-news_tonga_tonga_islands_speech_t5_summarizer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English news_tonga_tonga_islands_speech_t5_summarizer_pipeline pipeline T5Transformer from wrongserenity +author: John Snow Labs +name: news_tonga_tonga_islands_speech_t5_summarizer_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`news_tonga_tonga_islands_speech_t5_summarizer_pipeline` is a English model originally trained by wrongserenity. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/news_tonga_tonga_islands_speech_t5_summarizer_pipeline_en_5.4.2_3.0_1723916606786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/news_tonga_tonga_islands_speech_t5_summarizer_pipeline_en_5.4.2_3.0_1723916606786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("news_tonga_tonga_islands_speech_t5_summarizer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("news_tonga_tonga_islands_speech_t5_summarizer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|news_tonga_tonga_islands_speech_t5_summarizer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|976.8 MB| + +## References + +https://huggingface.co/wrongserenity/news_to_speech_t5_summarizer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_en.md new file mode 100644 index 00000000000000..f4521b9b245750 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English poem_gen_t5_small T5Transformer from DrishtiSharma +author: John Snow Labs +name: poem_gen_t5_small +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poem_gen_t5_small` is a English model originally trained by DrishtiSharma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_en_5.4.2_3.0_1723929026280.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_en_5.4.2_3.0_1723929026280.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("poem_gen_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("poem_gen_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poem_gen_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|327.7 MB| + +## References + +https://huggingface.co/DrishtiSharma/poem-gen-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..10234484790be4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-poem_gen_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English poem_gen_t5_small_pipeline pipeline T5Transformer from DrishtiSharma +author: John Snow Labs +name: poem_gen_t5_small_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poem_gen_t5_small_pipeline` is a English model originally trained by DrishtiSharma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_pipeline_en_5.4.2_3.0_1723929049596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_pipeline_en_5.4.2_3.0_1723929049596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("poem_gen_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("poem_gen_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poem_gen_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|327.7 MB| + +## References + +https://huggingface.co/DrishtiSharma/poem-gen-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_en.md b/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_en.md new file mode 100644 index 00000000000000..1c84723f67176d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English polytao_bigsmiles_version T5Transformer from hkqiu +author: John Snow Labs +name: polytao_bigsmiles_version +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polytao_bigsmiles_version` is a English model originally trained by hkqiu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polytao_bigsmiles_version_en_5.4.2_3.0_1723920598037.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polytao_bigsmiles_version_en_5.4.2_3.0_1723920598037.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("polytao_bigsmiles_version","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("polytao_bigsmiles_version", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polytao_bigsmiles_version| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hkqiu/PolyTAO-BigSMILES_Version \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_pipeline_en.md new file mode 100644 index 00000000000000..8c3fe79fa1c8f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-polytao_bigsmiles_version_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English polytao_bigsmiles_version_pipeline pipeline T5Transformer from hkqiu +author: John Snow Labs +name: polytao_bigsmiles_version_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polytao_bigsmiles_version_pipeline` is a English model originally trained by hkqiu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polytao_bigsmiles_version_pipeline_en_5.4.2_3.0_1723920648141.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polytao_bigsmiles_version_pipeline_en_5.4.2_3.0_1723920648141.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("polytao_bigsmiles_version_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("polytao_bigsmiles_version_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polytao_bigsmiles_version_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hkqiu/PolyTAO-BigSMILES_Version + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pipeline_pt.md b/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pipeline_pt.md new file mode 100644 index 00000000000000..e3bc690975e90c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pipeline_pt.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Portuguese ptt5_base_english_portuguese_msmarco_100k_v2_pipeline pipeline T5Transformer from unicamp-dl +author: John Snow Labs +name: ptt5_base_english_portuguese_msmarco_100k_v2_pipeline +date: 2024-08-17 +tags: [pt, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_base_english_portuguese_msmarco_100k_v2_pipeline` is a Portuguese model originally trained by unicamp-dl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_base_english_portuguese_msmarco_100k_v2_pipeline_pt_5.4.2_3.0_1723913782533.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_base_english_portuguese_msmarco_100k_v2_pipeline_pt_5.4.2_3.0_1723913782533.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ptt5_base_english_portuguese_msmarco_100k_v2_pipeline", lang = "pt") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ptt5_base_english_portuguese_msmarco_100k_v2_pipeline", lang = "pt") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_base_english_portuguese_msmarco_100k_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|pt| +|Size:|520.4 MB| + +## References + +https://huggingface.co/unicamp-dl/ptt5-base-en-pt-msmarco-100k-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pt.md b/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pt.md new file mode 100644 index 00000000000000..2b117f19b1cdca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-ptt5_base_english_portuguese_msmarco_100k_v2_pt.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Portuguese ptt5_base_english_portuguese_msmarco_100k_v2 T5Transformer from unicamp-dl +author: John Snow Labs +name: ptt5_base_english_portuguese_msmarco_100k_v2 +date: 2024-08-17 +tags: [pt, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_base_english_portuguese_msmarco_100k_v2` is a Portuguese model originally trained by unicamp-dl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_base_english_portuguese_msmarco_100k_v2_pt_5.4.2_3.0_1723913602562.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_base_english_portuguese_msmarco_100k_v2_pt_5.4.2_3.0_1723913602562.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ptt5_base_english_portuguese_msmarco_100k_v2","pt") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ptt5_base_english_portuguese_msmarco_100k_v2", "pt") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_base_english_portuguese_msmarco_100k_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|pt| +|Size:|520.4 MB| + +## References + +https://huggingface.co/unicamp-dl/ptt5-base-en-pt-msmarco-100k-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_en.md new file mode 100644 index 00000000000000..59b124a75b2a3c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qqp_t5_base_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: qqp_t5_base_seed_1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qqp_t5_base_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qqp_t5_base_seed_1_en_5.4.2_3.0_1723936857599.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qqp_t5_base_seed_1_en_5.4.2_3.0_1723936857599.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qqp_t5_base_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qqp_t5_base_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qqp_t5_base_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|980.5 MB| + +## References + +https://huggingface.co/utahnlp/qqp_t5-base_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..673ecf783c4bc8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-qqp_t5_base_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qqp_t5_base_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: qqp_t5_base_seed_1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qqp_t5_base_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qqp_t5_base_seed_1_pipeline_en_5.4.2_3.0_1723936915712.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qqp_t5_base_seed_1_pipeline_en_5.4.2_3.0_1723936915712.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qqp_t5_base_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qqp_t5_base_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qqp_t5_base_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|980.5 MB| + +## References + +https://huggingface.co/utahnlp/qqp_t5-base_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_en.md b/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_en.md new file mode 100644 index 00000000000000..a32146e4f6257f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qqq_finetuned_on_calls T5Transformer from Mauregato +author: John Snow Labs +name: qqq_finetuned_on_calls +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qqq_finetuned_on_calls` is a English model originally trained by Mauregato. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qqq_finetuned_on_calls_en_5.4.2_3.0_1723935670597.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qqq_finetuned_on_calls_en_5.4.2_3.0_1723935670597.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qqq_finetuned_on_calls","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qqq_finetuned_on_calls", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qqq_finetuned_on_calls| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|996.6 MB| + +## References + +https://huggingface.co/Mauregato/qqq-finetuned-on-calls \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_pipeline_en.md new file mode 100644 index 00000000000000..e63b9aea461ccf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-qqq_finetuned_on_calls_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qqq_finetuned_on_calls_pipeline pipeline T5Transformer from Mauregato +author: John Snow Labs +name: qqq_finetuned_on_calls_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qqq_finetuned_on_calls_pipeline` is a English model originally trained by Mauregato. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qqq_finetuned_on_calls_pipeline_en_5.4.2_3.0_1723935718667.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qqq_finetuned_on_calls_pipeline_en_5.4.2_3.0_1723935718667.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qqq_finetuned_on_calls_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qqq_finetuned_on_calls_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qqq_finetuned_on_calls_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|996.6 MB| + +## References + +https://huggingface.co/Mauregato/qqq-finetuned-on-calls + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_en.md new file mode 100644 index 00000000000000..4d3478fc67af31 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English real_prompt_300_500syn_root_cause_gen_t5_small T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_500syn_root_cause_gen_t5_small +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_500syn_root_cause_gen_t5_small` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_500syn_root_cause_gen_t5_small_en_5.4.2_3.0_1723912257676.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_500syn_root_cause_gen_t5_small_en_5.4.2_3.0_1723912257676.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("real_prompt_300_500syn_root_cause_gen_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("real_prompt_300_500syn_root_cause_gen_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_500syn_root_cause_gen_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-500syn-root_cause-gen-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..5189c5deb6077b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_500syn_root_cause_gen_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English real_prompt_300_500syn_root_cause_gen_t5_small_pipeline pipeline T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_500syn_root_cause_gen_t5_small_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_500syn_root_cause_gen_t5_small_pipeline` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_500syn_root_cause_gen_t5_small_pipeline_en_5.4.2_3.0_1723912321452.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_500syn_root_cause_gen_t5_small_pipeline_en_5.4.2_3.0_1723912321452.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("real_prompt_300_500syn_root_cause_gen_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("real_prompt_300_500syn_root_cause_gen_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_500syn_root_cause_gen_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-500syn-root_cause-gen-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_en.md new file mode 100644 index 00000000000000..f4c9a68c346d34 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English real_prompt_300_root_cause_gen_t5_small T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_root_cause_gen_t5_small +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_root_cause_gen_t5_small` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_root_cause_gen_t5_small_en_5.4.2_3.0_1723920635869.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_root_cause_gen_t5_small_en_5.4.2_3.0_1723920635869.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("real_prompt_300_root_cause_gen_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("real_prompt_300_root_cause_gen_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_root_cause_gen_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-root_cause-gen-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..00811eecfc5faf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-real_prompt_300_root_cause_gen_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English real_prompt_300_root_cause_gen_t5_small_pipeline pipeline T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_root_cause_gen_t5_small_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_root_cause_gen_t5_small_pipeline` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_root_cause_gen_t5_small_pipeline_en_5.4.2_3.0_1723920697285.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_root_cause_gen_t5_small_pipeline_en_5.4.2_3.0_1723920697285.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("real_prompt_300_root_cause_gen_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("real_prompt_300_root_cause_gen_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_root_cause_gen_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-root_cause-gen-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_bn.md b/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_bn.md new file mode 100644 index 00000000000000..92b3e99ecd5ea6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_bn.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Bengali regben2ipa_mt5_base T5Transformer from teamapocalypseml +author: John Snow Labs +name: regben2ipa_mt5_base +date: 2024-08-17 +tags: [bn, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`regben2ipa_mt5_base` is a Bengali model originally trained by teamapocalypseml. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/regben2ipa_mt5_base_bn_5.4.2_3.0_1723938013089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/regben2ipa_mt5_base_bn_5.4.2_3.0_1723938013089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("regben2ipa_mt5_base","bn") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("regben2ipa_mt5_base", "bn") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|regben2ipa_mt5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|bn| +|Size:|2.2 GB| + +## References + +https://huggingface.co/teamapocalypseml/regben2ipa-mt5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_pipeline_bn.md b/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_pipeline_bn.md new file mode 100644 index 00000000000000..832335cde632ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-regben2ipa_mt5_base_pipeline_bn.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Bengali regben2ipa_mt5_base_pipeline pipeline T5Transformer from teamapocalypseml +author: John Snow Labs +name: regben2ipa_mt5_base_pipeline +date: 2024-08-17 +tags: [bn, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`regben2ipa_mt5_base_pipeline` is a Bengali model originally trained by teamapocalypseml. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/regben2ipa_mt5_base_pipeline_bn_5.4.2_3.0_1723938178448.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/regben2ipa_mt5_base_pipeline_bn_5.4.2_3.0_1723938178448.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("regben2ipa_mt5_base_pipeline", lang = "bn") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("regben2ipa_mt5_base_pipeline", lang = "bn") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|regben2ipa_mt5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|bn| +|Size:|2.2 GB| + +## References + +https://huggingface.co/teamapocalypseml/regben2ipa-mt5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_en.md b/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_en.md new file mode 100644 index 00000000000000..f62f7f67b85eba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English results2_ara_t5_v2 T5Transformer from hiba2 +author: John Snow Labs +name: results2_ara_t5_v2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results2_ara_t5_v2` is a English model originally trained by hiba2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results2_ara_t5_v2_en_5.4.2_3.0_1723938772098.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results2_ara_t5_v2_en_5.4.2_3.0_1723938772098.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("results2_ara_t5_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("results2_ara_t5_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results2_ara_t5_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|972.2 MB| + +## References + +https://huggingface.co/hiba2/results2_ara_t5_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_pipeline_en.md new file mode 100644 index 00000000000000..fe2b36680b439a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-results2_ara_t5_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English results2_ara_t5_v2_pipeline pipeline T5Transformer from hiba2 +author: John Snow Labs +name: results2_ara_t5_v2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results2_ara_t5_v2_pipeline` is a English model originally trained by hiba2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results2_ara_t5_v2_pipeline_en_5.4.2_3.0_1723938840931.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results2_ara_t5_v2_pipeline_en_5.4.2_3.0_1723938840931.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("results2_ara_t5_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("results2_ara_t5_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results2_ara_t5_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|972.2 MB| + +## References + +https://huggingface.co/hiba2/results2_ara_t5_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_en.md b/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_en.md new file mode 100644 index 00000000000000..60c4d970a86c47 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English results_mt5_finetuned_squad_accelerate_m5 T5Transformer from YuTingHu +author: John Snow Labs +name: results_mt5_finetuned_squad_accelerate_m5 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_mt5_finetuned_squad_accelerate_m5` is a English model originally trained by YuTingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_mt5_finetuned_squad_accelerate_m5_en_5.4.2_3.0_1723930276696.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_mt5_finetuned_squad_accelerate_m5_en_5.4.2_3.0_1723930276696.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("results_mt5_finetuned_squad_accelerate_m5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("results_mt5_finetuned_squad_accelerate_m5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_mt5_finetuned_squad_accelerate_m5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/YuTingHu/results-mt5-finetuned-squad-accelerate_M5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_pipeline_en.md new file mode 100644 index 00000000000000..e9f5a52806e9fe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-results_mt5_finetuned_squad_accelerate_m5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English results_mt5_finetuned_squad_accelerate_m5_pipeline pipeline T5Transformer from YuTingHu +author: John Snow Labs +name: results_mt5_finetuned_squad_accelerate_m5_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_mt5_finetuned_squad_accelerate_m5_pipeline` is a English model originally trained by YuTingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_mt5_finetuned_squad_accelerate_m5_pipeline_en_5.4.2_3.0_1723930422288.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_mt5_finetuned_squad_accelerate_m5_pipeline_en_5.4.2_3.0_1723930422288.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("results_mt5_finetuned_squad_accelerate_m5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("results_mt5_finetuned_squad_accelerate_m5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_mt5_finetuned_squad_accelerate_m5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/YuTingHu/results-mt5-finetuned-squad-accelerate_M5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_en.md b/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_en.md new file mode 100644 index 00000000000000..6c35e52157e7b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rut5_summarizer_v2 T5Transformer from uselezzz +author: John Snow Labs +name: rut5_summarizer_v2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_summarizer_v2` is a English model originally trained by uselezzz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_summarizer_v2_en_5.4.2_3.0_1723935463989.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_summarizer_v2_en_5.4.2_3.0_1723935463989.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_summarizer_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_summarizer_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_summarizer_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/uselezzz/ruT5-summarizer-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_pipeline_en.md new file mode 100644 index 00000000000000..f9e763eb344900 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-rut5_summarizer_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rut5_summarizer_v2_pipeline pipeline T5Transformer from uselezzz +author: John Snow Labs +name: rut5_summarizer_v2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_summarizer_v2_pipeline` is a English model originally trained by uselezzz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_summarizer_v2_pipeline_en_5.4.2_3.0_1723935514102.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_summarizer_v2_pipeline_en_5.4.2_3.0_1723935514102.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_summarizer_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_summarizer_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_summarizer_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/uselezzz/ruT5-summarizer-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_en.md b/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_en.md new file mode 100644 index 00000000000000..1aef64a946b879 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3 T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_en_5.4.2_3.0_1723932767663.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_en_5.4.2_3.0_1723932767663.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|317.0 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline_en.md new file mode 100644 index 00000000000000..ce5c6990a149f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline_en_5.4.2_3.0_1723932785759.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline_en_5.4.2_3.0_1723932785759.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|317.0 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_en.md b/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_en.md new file mode 100644 index 00000000000000..6e97ab276c2773 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English single_sentence_models T5Transformer from sheoran95 +author: John Snow Labs +name: single_sentence_models +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`single_sentence_models` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/single_sentence_models_en_5.4.2_3.0_1723916262286.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/single_sentence_models_en_5.4.2_3.0_1723916262286.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("single_sentence_models","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("single_sentence_models", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|single_sentence_models| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|325.6 MB| + +## References + +https://huggingface.co/sheoran95/single_sentence_models \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_pipeline_en.md new file mode 100644 index 00000000000000..09783f2b92ab53 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-single_sentence_models_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English single_sentence_models_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: single_sentence_models_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`single_sentence_models_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/single_sentence_models_pipeline_en_5.4.2_3.0_1723916280563.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/single_sentence_models_pipeline_en_5.4.2_3.0_1723916280563.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("single_sentence_models_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("single_sentence_models_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|single_sentence_models_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|325.6 MB| + +## References + +https://huggingface.co/sheoran95/single_sentence_models + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_en.md b/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_en.md new file mode 100644 index 00000000000000..141362ac24130d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English software_req_module_identification T5Transformer from MeetK +author: John Snow Labs +name: software_req_module_identification +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`software_req_module_identification` is a English model originally trained by MeetK. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/software_req_module_identification_en_5.4.2_3.0_1723923855588.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/software_req_module_identification_en_5.4.2_3.0_1723923855588.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("software_req_module_identification","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("software_req_module_identification", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|software_req_module_identification| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|313.7 MB| + +## References + +https://huggingface.co/MeetK/software_req_module_identification \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_pipeline_en.md new file mode 100644 index 00000000000000..a3c0fac21ab186 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-software_req_module_identification_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English software_req_module_identification_pipeline pipeline T5Transformer from MeetK +author: John Snow Labs +name: software_req_module_identification_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`software_req_module_identification_pipeline` is a English model originally trained by MeetK. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/software_req_module_identification_pipeline_en_5.4.2_3.0_1723923878488.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/software_req_module_identification_pipeline_en_5.4.2_3.0_1723923878488.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("software_req_module_identification_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("software_req_module_identification_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|software_req_module_identification_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|313.7 MB| + +## References + +https://huggingface.co/MeetK/software_req_module_identification + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_en.md b/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_en.md new file mode 100644 index 00000000000000..0c15045dd83c41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English subreddit_thread_tagging T5Transformer from textomatic +author: John Snow Labs +name: subreddit_thread_tagging +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`subreddit_thread_tagging` is a English model originally trained by textomatic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/subreddit_thread_tagging_en_5.4.2_3.0_1723924419451.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/subreddit_thread_tagging_en_5.4.2_3.0_1723924419451.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("subreddit_thread_tagging","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("subreddit_thread_tagging", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|subreddit_thread_tagging| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/textomatic/subreddit-thread-tagging \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_pipeline_en.md new file mode 100644 index 00000000000000..cfe26cbea7b20c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-subreddit_thread_tagging_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English subreddit_thread_tagging_pipeline pipeline T5Transformer from textomatic +author: John Snow Labs +name: subreddit_thread_tagging_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`subreddit_thread_tagging_pipeline` is a English model originally trained by textomatic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/subreddit_thread_tagging_pipeline_en_5.4.2_3.0_1723924436444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/subreddit_thread_tagging_pipeline_en_5.4.2_3.0_1723924436444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("subreddit_thread_tagging_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("subreddit_thread_tagging_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|subreddit_thread_tagging_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/textomatic/subreddit-thread-tagging + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_en.md b/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_en.md new file mode 100644 index 00000000000000..5b0d63185d54a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarization_model_samuela39 T5Transformer from Samuela39 +author: John Snow Labs +name: summarization_model_samuela39 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_model_samuela39` is a English model originally trained by Samuela39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_model_samuela39_en_5.4.2_3.0_1723920225097.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_model_samuela39_en_5.4.2_3.0_1723920225097.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarization_model_samuela39","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarization_model_samuela39", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_model_samuela39| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|323.4 MB| + +## References + +https://huggingface.co/Samuela39/summarization_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_pipeline_en.md new file mode 100644 index 00000000000000..ab39a893057764 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-summarization_model_samuela39_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarization_model_samuela39_pipeline pipeline T5Transformer from Samuela39 +author: John Snow Labs +name: summarization_model_samuela39_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_model_samuela39_pipeline` is a English model originally trained by Samuela39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_model_samuela39_pipeline_en_5.4.2_3.0_1723920247738.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_model_samuela39_pipeline_en_5.4.2_3.0_1723920247738.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarization_model_samuela39_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarization_model_samuela39_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_model_samuela39_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|323.4 MB| + +## References + +https://huggingface.co/Samuela39/summarization_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_en.md new file mode 100644 index 00000000000000..11ef2abe17f292 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_aic_combined_years_with_year_flag T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_aic_combined_years_with_year_flag +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_aic_combined_years_with_year_flag` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_aic_combined_years_with_year_flag_en_5.4.2_3.0_1723919732077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_aic_combined_years_with_year_flag_en_5.4.2_3.0_1723919732077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_aic_combined_years_with_year_flag","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_aic_combined_years_with_year_flag", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_aic_combined_years_with_year_flag| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|322.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-aic-combined_years_with_year_flag \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_pipeline_en.md new file mode 100644 index 00000000000000..dcf4da5cd89aff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_aic_combined_years_with_year_flag_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_aic_combined_years_with_year_flag_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_aic_combined_years_with_year_flag_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_aic_combined_years_with_year_flag_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_aic_combined_years_with_year_flag_pipeline_en_5.4.2_3.0_1723919755856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_aic_combined_years_with_year_flag_pipeline_en_5.4.2_3.0_1723919755856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_aic_combined_years_with_year_flag_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_aic_combined_years_with_year_flag_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_aic_combined_years_with_year_flag_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|322.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-aic-combined_years_with_year_flag + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_en.md new file mode 100644 index 00000000000000..c90366cd49516a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_4 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_4 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_4` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_4_en_5.4.2_3.0_1723920379025.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_4_en_5.4.2_3.0_1723920379025.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_pipeline_en.md new file mode 100644 index 00000000000000..b06cb77b72eb9d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_60m_lm_wmt_2018_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_4_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_4_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_4_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_4_pipeline_en_5.4.2_3.0_1723920396551.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_4_pipeline_en_5.4.2_3.0_1723920396551.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2018_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2018_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_770m_pacit_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_770m_pacit_en.md new file mode 100644 index 00000000000000..c0013e4df452f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_770m_pacit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_770m_pacit T5Transformer from happy-frog +author: John Snow Labs +name: t5_770m_pacit +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_770m_pacit` is a English model originally trained by happy-frog. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_770m_pacit_en_5.4.2_3.0_1723914814720.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_770m_pacit_en_5.4.2_3.0_1723914814720.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_770m_pacit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_770m_pacit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_770m_pacit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/happy-frog/T5_770M-PACIT \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_en.md new file mode 100644 index 00000000000000..9b7875fb29a2a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_billsum_dense_epochs_6 T5Transformer from thrunlab +author: John Snow Labs +name: t5_base_billsum_dense_epochs_6 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_billsum_dense_epochs_6` is a English model originally trained by thrunlab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_billsum_dense_epochs_6_en_5.4.2_3.0_1723927458218.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_billsum_dense_epochs_6_en_5.4.2_3.0_1723927458218.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_billsum_dense_epochs_6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_billsum_dense_epochs_6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_billsum_dense_epochs_6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|992.9 MB| + +## References + +https://huggingface.co/thrunlab/t5-base_billsum_dense_epochs-6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_pipeline_en.md new file mode 100644 index 00000000000000..51cc3e3582b0ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_billsum_dense_epochs_6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_billsum_dense_epochs_6_pipeline pipeline T5Transformer from thrunlab +author: John Snow Labs +name: t5_base_billsum_dense_epochs_6_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_billsum_dense_epochs_6_pipeline` is a English model originally trained by thrunlab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_billsum_dense_epochs_6_pipeline_en_5.4.2_3.0_1723927511875.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_billsum_dense_epochs_6_pipeline_en_5.4.2_3.0_1723927511875.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_billsum_dense_epochs_6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_billsum_dense_epochs_6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_billsum_dense_epochs_6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|992.9 MB| + +## References + +https://huggingface.co/thrunlab/t5-base_billsum_dense_epochs-6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_en.md new file mode 100644 index 00000000000000..4f6f6c953984ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt2_khanq T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt2_khanq +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt2_khanq` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_en_5.4.2_3.0_1723938864704.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_en_5.4.2_3.0_1723938864704.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt2_khanq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt2_khanq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt2_khanq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt2-khanq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_pipeline_en.md new file mode 100644 index 00000000000000..d6e521ec7d8fff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt2_khanq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt2_khanq_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt2_khanq_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt2_khanq_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_pipeline_en_5.4.2_3.0_1723938916892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_pipeline_en_5.4.2_3.0_1723938916892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt2_khanq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt2_khanq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt2_khanq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt2-khanq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_en.md new file mode 100644 index 00000000000000..f31d77ecb831d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt3_khanq T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt3_khanq +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt3_khanq` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt3_khanq_en_5.4.2_3.0_1723916401231.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt3_khanq_en_5.4.2_3.0_1723916401231.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt3_khanq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt3_khanq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt3_khanq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt3-khanq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_pipeline_en.md new file mode 100644 index 00000000000000..20f49c8c65e803 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_bt3_khanq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt3_khanq_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt3_khanq_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt3_khanq_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt3_khanq_pipeline_en_5.4.2_3.0_1723916451368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt3_khanq_pipeline_en_5.4.2_3.0_1723916451368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt3_khanq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt3_khanq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt3_khanq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt3-khanq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_en.md new file mode 100644 index 00000000000000..001d4124096177 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_en_5.4.2_3.0_1723911744186.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_en_5.4.2_3.0_1723911744186.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|934.3 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-32-finetuned-squad-infilling-seed-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline_en.md new file mode 100644 index 00000000000000..4240bd6af13e7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline_en_5.4.2_3.0_1723911823309.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline_en_5.4.2_3.0_1723911823309.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_32_finetuned_squad_infilling_seed_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|934.3 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-32-finetuned-squad-infilling-seed-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_en.md new file mode 100644 index 00000000000000..cc91aefc574e49 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1723933262332.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1723933262332.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|934.4 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-32-finetuned-squad-infilling-seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..f5693e9f5d9bfc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1723933338384.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1723933338384.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_32_finetuned_squad_infilling_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|934.4 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-32-finetuned-squad-infilling-seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_en.md new file mode 100644 index 00000000000000..1109885282d052 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_bbc_lemmatized T5Transformer from piyushjain4 +author: John Snow Labs +name: t5_base_finetuned_bbc_lemmatized +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_bbc_lemmatized` is a English model originally trained by piyushjain4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_lemmatized_en_5.4.2_3.0_1723912956638.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_lemmatized_en_5.4.2_3.0_1723912956638.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_bbc_lemmatized","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_bbc_lemmatized", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_bbc_lemmatized| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|988.5 MB| + +## References + +https://huggingface.co/piyushjain4/t5-base-finetuned-bbc-lemmatized \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_pipeline_en.md new file mode 100644 index 00000000000000..7d59867252d6f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_finetuned_bbc_lemmatized_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_bbc_lemmatized_pipeline pipeline T5Transformer from piyushjain4 +author: John Snow Labs +name: t5_base_finetuned_bbc_lemmatized_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_bbc_lemmatized_pipeline` is a English model originally trained by piyushjain4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_lemmatized_pipeline_en_5.4.2_3.0_1723913007676.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_lemmatized_pipeline_en_5.4.2_3.0_1723913007676.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_bbc_lemmatized_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_bbc_lemmatized_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_bbc_lemmatized_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|988.5 MB| + +## References + +https://huggingface.co/piyushjain4/t5-base-finetuned-bbc-lemmatized + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_en.md new file mode 100644 index 00000000000000..b743a0d0a78098 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_mediasum_seed42 T5Transformer from PSW +author: John Snow Labs +name: t5_base_mediasum_seed42 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_mediasum_seed42` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed42_en_5.4.2_3.0_1723936354810.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed42_en_5.4.2_3.0_1723936354810.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_mediasum_seed42","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_mediasum_seed42", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_mediasum_seed42| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-mediasum-seed42 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_pipeline_en.md new file mode 100644 index 00000000000000..08444be6a7e4f5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_mediasum_seed42_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_mediasum_seed42_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_mediasum_seed42_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_mediasum_seed42_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed42_pipeline_en_5.4.2_3.0_1723936410470.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed42_pipeline_en_5.4.2_3.0_1723936410470.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_mediasum_seed42_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_mediasum_seed42_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_mediasum_seed42_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-mediasum-seed42 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_en.md new file mode 100644 index 00000000000000..5da108d4100e74 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_movies T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_movies +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_movies` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_movies_en_5.4.2_3.0_1723917562868.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_movies_en_5.4.2_3.0_1723917562868.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_movies","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_movies", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_movies| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|977.8 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-movies \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_pipeline_en.md new file mode 100644 index 00000000000000..76aa10235e631e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_rlhf_bm25_movies_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_movies_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_movies_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_movies_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_movies_pipeline_en_5.4.2_3.0_1723917619901.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_movies_pipeline_en_5.4.2_3.0_1723917619901.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_bm25_movies_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_bm25_movies_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_movies_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|977.8 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-movies + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_en.md new file mode 100644 index 00000000000000..e664560bcfb4a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_8front_1body_0rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_8front_1body_0rear +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_8front_1body_0rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_8front_1body_0rear_en_5.4.2_3.0_1723926709191.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_8front_1body_0rear_en_5.4.2_3.0_1723926709191.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_8front_1body_0rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_8front_1body_0rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_8front_1body_0rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-8front-1body-0rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_pipeline_en.md new file mode 100644 index 00000000000000..ba195f41f41b91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_tedxjp_8front_1body_0rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_8front_1body_0rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_8front_1body_0rear_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_8front_1body_0rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_8front_1body_0rear_pipeline_en_5.4.2_3.0_1723926760467.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_8front_1body_0rear_pipeline_en_5.4.2_3.0_1723926760467.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_8front_1body_0rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_8front_1body_0rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_8front_1body_0rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-8front-1body-0rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_en.md new file mode 100644 index 00000000000000..934965c9657481 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_trained_vishal T5Transformer from vishalgupta +author: John Snow Labs +name: t5_base_trained_vishal +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_trained_vishal` is a English model originally trained by vishalgupta. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_trained_vishal_en_5.4.2_3.0_1723912846798.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_trained_vishal_en_5.4.2_3.0_1723912846798.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_trained_vishal","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_trained_vishal", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_trained_vishal| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|965.8 MB| + +## References + +https://huggingface.co/vishalgupta/t5-base-trained-vishal \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_pipeline_en.md new file mode 100644 index 00000000000000..a7ae166e467854 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_base_trained_vishal_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_trained_vishal_pipeline pipeline T5Transformer from vishalgupta +author: John Snow Labs +name: t5_base_trained_vishal_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_trained_vishal_pipeline` is a English model originally trained by vishalgupta. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_trained_vishal_pipeline_en_5.4.2_3.0_1723912906628.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_trained_vishal_pipeline_en_5.4.2_3.0_1723912906628.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_trained_vishal_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_trained_vishal_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_trained_vishal_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|965.8 MB| + +## References + +https://huggingface.co/vishalgupta/t5-base-trained-vishal + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_en.md new file mode 100644 index 00000000000000..42113942921af3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_booksum T5Transformer from saketh092 +author: John Snow Labs +name: t5_booksum +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_booksum` is a English model originally trained by saketh092. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_booksum_en_5.4.2_3.0_1723925451376.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_booksum_en_5.4.2_3.0_1723925451376.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_booksum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_booksum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_booksum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|183.6 MB| + +## References + +https://huggingface.co/saketh092/t5-booksum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_pipeline_en.md new file mode 100644 index 00000000000000..3c23a4dabcd854 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_booksum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_booksum_pipeline pipeline T5Transformer from saketh092 +author: John Snow Labs +name: t5_booksum_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_booksum_pipeline` is a English model originally trained by saketh092. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_booksum_pipeline_en_5.4.2_3.0_1723925513077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_booksum_pipeline_en_5.4.2_3.0_1723925513077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_booksum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_booksum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_booksum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|183.6 MB| + +## References + +https://huggingface.co/saketh092/t5-booksum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_en.md new file mode 100644 index 00000000000000..3ffa14c3902162 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_5 T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_5 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_5` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_5_en_5.4.2_3.0_1723912708192.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_5_en_5.4.2_3.0_1723912708192.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|988.7 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_pipeline_en.md new file mode 100644 index 00000000000000..bcc41f362abe4f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_e2e_10epochs_lr1e4_alpha0_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_5_pipeline pipeline T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_5_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_5_pipeline` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_5_pipeline_en_5.4.2_3.0_1723912762974.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_5_pipeline_en_5.4.2_3.0_1723912762974.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|988.7 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_en.md new file mode 100644 index 00000000000000..9a03ba2837bbb7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_ep3_parabk2 T5Transformer from jinujinu99 +author: John Snow Labs +name: t5_ep3_parabk2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ep3_parabk2` is a English model originally trained by jinujinu99. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ep3_parabk2_en_5.4.2_3.0_1723911618923.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ep3_parabk2_en_5.4.2_3.0_1723911618923.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_ep3_parabk2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_ep3_parabk2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ep3_parabk2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jinujinu99/t5-ep3-parabk2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_pipeline_en.md new file mode 100644 index 00000000000000..670b9861a8d59b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_ep3_parabk2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_ep3_parabk2_pipeline pipeline T5Transformer from jinujinu99 +author: John Snow Labs +name: t5_ep3_parabk2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ep3_parabk2_pipeline` is a English model originally trained by jinujinu99. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ep3_parabk2_pipeline_en_5.4.2_3.0_1723911666718.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ep3_parabk2_pipeline_en_5.4.2_3.0_1723911666718.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_ep3_parabk2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_ep3_parabk2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ep3_parabk2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jinujinu99/t5-ep3-parabk2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_en.md new file mode 100644 index 00000000000000..75cdb2341dfa1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_medical_english_german T5Transformer from muibk +author: John Snow Labs +name: t5_finetuned_medical_english_german +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_medical_english_german` is a English model originally trained by muibk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_medical_english_german_en_5.4.2_3.0_1723910343956.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_medical_english_german_en_5.4.2_3.0_1723910343956.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_medical_english_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_medical_english_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_medical_english_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|318.1 MB| + +## References + +https://huggingface.co/muibk/t5_finetuned_medical_en-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_pipeline_en.md new file mode 100644 index 00000000000000..12d108f32d9fed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_finetuned_medical_english_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_medical_english_german_pipeline pipeline T5Transformer from muibk +author: John Snow Labs +name: t5_finetuned_medical_english_german_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_medical_english_german_pipeline` is a English model originally trained by muibk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_medical_english_german_pipeline_en_5.4.2_3.0_1723910364327.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_medical_english_german_pipeline_en_5.4.2_3.0_1723910364327.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_medical_english_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_medical_english_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_medical_english_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|318.1 MB| + +## References + +https://huggingface.co/muibk/t5_finetuned_medical_en-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_en.md new file mode 100644 index 00000000000000..22f139e4269175 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_gec_128len_6e T5Transformer from anujraymajhi +author: John Snow Labs +name: t5_gec_128len_6e +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec_128len_6e` is a English model originally trained by anujraymajhi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_128len_6e_en_5.4.2_3.0_1723915518266.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_128len_6e_en_5.4.2_3.0_1723915518266.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_gec_128len_6e","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_gec_128len_6e", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec_128len_6e| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|996.3 MB| + +## References + +https://huggingface.co/anujraymajhi/t5-GEC-128len-6e \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_pipeline_en.md new file mode 100644 index 00000000000000..7f429bacf7a4a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_gec_128len_6e_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_gec_128len_6e_pipeline pipeline T5Transformer from anujraymajhi +author: John Snow Labs +name: t5_gec_128len_6e_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec_128len_6e_pipeline` is a English model originally trained by anujraymajhi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_128len_6e_pipeline_en_5.4.2_3.0_1723915570709.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_128len_6e_pipeline_en_5.4.2_3.0_1723915570709.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_gec_128len_6e_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_gec_128len_6e_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec_128len_6e_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|996.3 MB| + +## References + +https://huggingface.co/anujraymajhi/t5-GEC-128len-6e + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_en.md new file mode 100644 index 00000000000000..923cfe5a9ed19e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_jstilb T5Transformer from jstilb +author: John Snow Labs +name: t5_jstilb +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_jstilb` is a English model originally trained by jstilb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jstilb_en_5.4.2_3.0_1723927307960.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jstilb_en_5.4.2_3.0_1723927307960.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_jstilb","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_jstilb", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jstilb| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|320.8 MB| + +## References + +https://huggingface.co/jstilb/t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_pipeline_en.md new file mode 100644 index 00000000000000..f841417ddc396e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_jstilb_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_jstilb_pipeline pipeline T5Transformer from jstilb +author: John Snow Labs +name: t5_jstilb_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_jstilb_pipeline` is a English model originally trained by jstilb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jstilb_pipeline_en_5.4.2_3.0_1723927328084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jstilb_pipeline_en_5.4.2_3.0_1723927328084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_jstilb_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_jstilb_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jstilb_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|320.8 MB| + +## References + +https://huggingface.co/jstilb/t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_large_fine_tuned_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_large_fine_tuned_en.md new file mode 100644 index 00000000000000..c1887378379494 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_large_fine_tuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_fine_tuned T5Transformer from rob06 +author: John Snow Labs +name: t5_large_fine_tuned +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_fine_tuned` is a English model originally trained by rob06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_fine_tuned_en_5.4.2_3.0_1723922115648.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_fine_tuned_en_5.4.2_3.0_1723922115648.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_fine_tuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_fine_tuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_fine_tuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/rob06/t5-large-fine-tuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_en.md new file mode 100644 index 00000000000000..f495373c8e96f5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_finetuned2 T5Transformer from kowsiknd +author: John Snow Labs +name: t5_large_finetuned2 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_finetuned2` is a English model originally trained by kowsiknd. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_finetuned2_en_5.4.2_3.0_1723909907570.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_finetuned2_en_5.4.2_3.0_1723909907570.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_finetuned2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_finetuned2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_finetuned2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/kowsiknd/t5-large-finetuned2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_pipeline_en.md new file mode 100644 index 00000000000000..f1058d05f21006 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_large_finetuned2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_finetuned2_pipeline pipeline T5Transformer from kowsiknd +author: John Snow Labs +name: t5_large_finetuned2_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_finetuned2_pipeline` is a English model originally trained by kowsiknd. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_finetuned2_pipeline_en_5.4.2_3.0_1723910044868.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_finetuned2_pipeline_en_5.4.2_3.0_1723910044868.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_finetuned2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_finetuned2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_finetuned2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/kowsiknd/t5-large-finetuned2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_en.md new file mode 100644 index 00000000000000..88b2d27ab95c97 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_make_natural T5Transformer from jaekwanyda +author: John Snow Labs +name: t5_large_make_natural +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_make_natural` is a English model originally trained by jaekwanyda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_make_natural_en_5.4.2_3.0_1723922523670.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_make_natural_en_5.4.2_3.0_1723922523670.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_make_natural","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_make_natural", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_make_natural| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jaekwanyda/T5_large_make_natural \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_pipeline_en.md new file mode 100644 index 00000000000000..6925b08198625a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_large_make_natural_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_make_natural_pipeline pipeline T5Transformer from jaekwanyda +author: John Snow Labs +name: t5_large_make_natural_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_make_natural_pipeline` is a English model originally trained by jaekwanyda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_make_natural_pipeline_en_5.4.2_3.0_1723922671035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_make_natural_pipeline_en_5.4.2_3.0_1723922671035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_make_natural_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_make_natural_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_make_natural_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jaekwanyda/T5_large_make_natural + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_pst_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_pst_en.md new file mode 100644 index 00000000000000..193719a6de9077 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_pst_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_pst T5Transformer from zaidbhatti +author: John Snow Labs +name: t5_pst +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pst` is a English model originally trained by zaidbhatti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pst_en_5.4.2_3.0_1723926131300.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pst_en_5.4.2_3.0_1723926131300.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_pst","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_pst", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pst| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|288.9 MB| + +## References + +https://huggingface.co/zaidbhatti/t5-pst \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_pst_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_pst_pipeline_en.md new file mode 100644 index 00000000000000..82cee75d3bc8b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_pst_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_pst_pipeline pipeline T5Transformer from zaidbhatti +author: John Snow Labs +name: t5_pst_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pst_pipeline` is a English model originally trained by zaidbhatti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pst_pipeline_en_5.4.2_3.0_1723926160268.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pst_pipeline_en_5.4.2_3.0_1723926160268.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_pst_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_pst_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pst_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|288.9 MB| + +## References + +https://huggingface.co/zaidbhatti/t5-pst + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_en.md new file mode 100644 index 00000000000000..6b3ddbb02fbc03 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_512 T5Transformer from bangnbx +author: John Snow Labs +name: t5_small_512 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_512` is a English model originally trained by bangnbx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_512_en_5.4.2_3.0_1723912931353.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_512_en_5.4.2_3.0_1723912931353.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_512","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_512", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_512| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/bangnbx/t5-small-512 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_pipeline_en.md new file mode 100644 index 00000000000000..d6293af7967978 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_512_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_512_pipeline pipeline T5Transformer from bangnbx +author: John Snow Labs +name: t5_small_512_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_512_pipeline` is a English model originally trained by bangnbx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_512_pipeline_en_5.4.2_3.0_1723912948314.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_512_pipeline_en_5.4.2_3.0_1723912948314.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_512_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_512_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_512_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/bangnbx/t5-small-512 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_en.md new file mode 100644 index 00000000000000..ec42107161df82 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_eli5_kk4real T5Transformer from kk4real +author: John Snow Labs +name: t5_small_finetuned_eli5_kk4real +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_eli5_kk4real` is a English model originally trained by kk4real. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_eli5_kk4real_en_5.4.2_3.0_1723936351881.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_eli5_kk4real_en_5.4.2_3.0_1723936351881.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_eli5_kk4real","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_eli5_kk4real", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_eli5_kk4real| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|341.2 MB| + +## References + +https://huggingface.co/kk4real/t5-small-finetuned-eli5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_pipeline_en.md new file mode 100644 index 00000000000000..209533baccd445 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_eli5_kk4real_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_eli5_kk4real_pipeline pipeline T5Transformer from kk4real +author: John Snow Labs +name: t5_small_finetuned_eli5_kk4real_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_eli5_kk4real_pipeline` is a English model originally trained by kk4real. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_eli5_kk4real_pipeline_en_5.4.2_3.0_1723936369074.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_eli5_kk4real_pipeline_en_5.4.2_3.0_1723936369074.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_eli5_kk4real_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_eli5_kk4real_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_eli5_kk4real_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|341.2 MB| + +## References + +https://huggingface.co/kk4real/t5-small-finetuned-eli5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_en.md new file mode 100644 index 00000000000000..3d428386701fb2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_venkatarajendra T5Transformer from venkatarajendra +author: John Snow Labs +name: t5_small_finetuned_xsum_venkatarajendra +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_venkatarajendra` is a English model originally trained by venkatarajendra. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_venkatarajendra_en_5.4.2_3.0_1723925714909.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_venkatarajendra_en_5.4.2_3.0_1723925714909.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_venkatarajendra","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_venkatarajendra", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_venkatarajendra| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.9 MB| + +## References + +https://huggingface.co/venkatarajendra/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_pipeline_en.md new file mode 100644 index 00000000000000..884abd14817aff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_venkatarajendra_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_venkatarajendra_pipeline pipeline T5Transformer from venkatarajendra +author: John Snow Labs +name: t5_small_finetuned_xsum_venkatarajendra_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_venkatarajendra_pipeline` is a English model originally trained by venkatarajendra. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_venkatarajendra_pipeline_en_5.4.2_3.0_1723925733377.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_venkatarajendra_pipeline_en_5.4.2_3.0_1723925733377.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_venkatarajendra_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_venkatarajendra_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_venkatarajendra_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.9 MB| + +## References + +https://huggingface.co/venkatarajendra/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_en.md new file mode 100644 index 00000000000000..d39b4a101a5091 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_yogeshchandrasekharuni T5Transformer from yogeshchandrasekharuni +author: John Snow Labs +name: t5_small_finetuned_xsum_yogeshchandrasekharuni +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_yogeshchandrasekharuni` is a English model originally trained by yogeshchandrasekharuni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_yogeshchandrasekharuni_en_5.4.2_3.0_1723921161033.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_yogeshchandrasekharuni_en_5.4.2_3.0_1723921161033.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_yogeshchandrasekharuni","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_yogeshchandrasekharuni", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_yogeshchandrasekharuni| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|265.6 MB| + +## References + +https://huggingface.co/yogeshchandrasekharuni/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline_en.md new file mode 100644 index 00000000000000..c3c106f6e7f8b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline pipeline T5Transformer from yogeshchandrasekharuni +author: John Snow Labs +name: t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline` is a English model originally trained by yogeshchandrasekharuni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline_en_5.4.2_3.0_1723921189487.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline_en_5.4.2_3.0_1723921189487.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_yogeshchandrasekharuni_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|265.6 MB| + +## References + +https://huggingface.co/yogeshchandrasekharuni/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_en.md new file mode 100644 index 00000000000000..c2b4c90c4cb570 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_pointer_cstop_artificial T5Transformer from WillHeld +author: John Snow Labs +name: t5_small_pointer_cstop_artificial +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_pointer_cstop_artificial` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_pointer_cstop_artificial_en_5.4.2_3.0_1723919901124.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_pointer_cstop_artificial_en_5.4.2_3.0_1723919901124.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_pointer_cstop_artificial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_pointer_cstop_artificial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_pointer_cstop_artificial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/WillHeld/t5-small-pointer-cstop_artificial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_pipeline_en.md new file mode 100644 index 00000000000000..b6041b14617729 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_pointer_cstop_artificial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_pointer_cstop_artificial_pipeline pipeline T5Transformer from WillHeld +author: John Snow Labs +name: t5_small_pointer_cstop_artificial_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_pointer_cstop_artificial_pipeline` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_pointer_cstop_artificial_pipeline_en_5.4.2_3.0_1723920103817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_pointer_cstop_artificial_pipeline_en_5.4.2_3.0_1723920103817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_pointer_cstop_artificial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_pointer_cstop_artificial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_pointer_cstop_artificial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/WillHeld/t5-small-pointer-cstop_artificial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_en.md new file mode 100644 index 00000000000000..f0e8e7b93ca6a9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_squad_qg_default T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squad_qg_default +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_qg_default` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_default_en_5.4.2_3.0_1723920886085.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_default_en_5.4.2_3.0_1723920886085.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_squad_qg_default","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_squad_qg_default", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_qg_default| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.2 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squad-qg-default \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_pipeline_en.md new file mode 100644 index 00000000000000..df177955d37267 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_squad_qg_default_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_squad_qg_default_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squad_qg_default_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_qg_default_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_default_pipeline_en_5.4.2_3.0_1723920904316.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_default_pipeline_en_5.4.2_3.0_1723920904316.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_squad_qg_default_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_squad_qg_default_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_qg_default_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.2 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squad-qg-default + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_en.md new file mode 100644 index 00000000000000..f478c72a7f03ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_stereoset_finetuned_hbrpoi T5Transformer from henryscheible +author: John Snow Labs +name: t5_small_stereoset_finetuned_hbrpoi +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_stereoset_finetuned_hbrpoi` is a English model originally trained by henryscheible. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_stereoset_finetuned_hbrpoi_en_5.4.2_3.0_1723937037909.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_stereoset_finetuned_hbrpoi_en_5.4.2_3.0_1723937037909.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_stereoset_finetuned_hbrpoi","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_stereoset_finetuned_hbrpoi", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_stereoset_finetuned_hbrpoi| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|307.0 MB| + +## References + +https://huggingface.co/henryscheible/t5-small_stereoset_finetuned_HBRPOI \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_pipeline_en.md new file mode 100644 index 00000000000000..c6774fe3f93825 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_small_stereoset_finetuned_hbrpoi_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_stereoset_finetuned_hbrpoi_pipeline pipeline T5Transformer from henryscheible +author: John Snow Labs +name: t5_small_stereoset_finetuned_hbrpoi_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_stereoset_finetuned_hbrpoi_pipeline` is a English model originally trained by henryscheible. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_stereoset_finetuned_hbrpoi_pipeline_en_5.4.2_3.0_1723937065023.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_stereoset_finetuned_hbrpoi_pipeline_en_5.4.2_3.0_1723937065023.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_stereoset_finetuned_hbrpoi_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_stereoset_finetuned_hbrpoi_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_stereoset_finetuned_hbrpoi_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|307.0 MB| + +## References + +https://huggingface.co/henryscheible/t5-small_stereoset_finetuned_HBRPOI + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_en.md new file mode 100644 index 00000000000000..0e59e629c1bb98 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_summarization_zero_shot_headers_and_better_prompt T5Transformer from veronica-girolimetti +author: John Snow Labs +name: t5_summarization_zero_shot_headers_and_better_prompt +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarization_zero_shot_headers_and_better_prompt` is a English model originally trained by veronica-girolimetti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_en_5.4.2_3.0_1723913046911.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_en_5.4.2_3.0_1723913046911.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_summarization_zero_shot_headers_and_better_prompt","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_summarization_zero_shot_headers_and_better_prompt", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarization_zero_shot_headers_and_better_prompt| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/veronica-girolimetti/t5-summarization-zero-shot-headers-and-better-prompt \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_pipeline_en.md new file mode 100644 index 00000000000000..38dcb75f46f682 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_summarization_zero_shot_headers_and_better_prompt_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_summarization_zero_shot_headers_and_better_prompt_pipeline pipeline T5Transformer from veronica-girolimetti +author: John Snow Labs +name: t5_summarization_zero_shot_headers_and_better_prompt_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarization_zero_shot_headers_and_better_prompt_pipeline` is a English model originally trained by veronica-girolimetti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_pipeline_en_5.4.2_3.0_1723913064065.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_pipeline_en_5.4.2_3.0_1723913064065.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_summarization_zero_shot_headers_and_better_prompt_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_summarization_zero_shot_headers_and_better_prompt_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarization_zero_shot_headers_and_better_prompt_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/veronica-girolimetti/t5-summarization-zero-shot-headers-and-better-prompt + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_en.md new file mode 100644 index 00000000000000..f3b126ddb5f3ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_textdetoxification T5Transformer from narySt +author: John Snow Labs +name: t5_textdetoxification +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_textdetoxification` is a English model originally trained by narySt. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_textdetoxification_en_5.4.2_3.0_1723914254680.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_textdetoxification_en_5.4.2_3.0_1723914254680.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_textdetoxification","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_textdetoxification", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_textdetoxification| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/narySt/T5_textDetoxification \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_pipeline_en.md new file mode 100644 index 00000000000000..31a4253253558d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_textdetoxification_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_textdetoxification_pipeline pipeline T5Transformer from narySt +author: John Snow Labs +name: t5_textdetoxification_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_textdetoxification_pipeline` is a English model originally trained by narySt. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_textdetoxification_pipeline_en_5.4.2_3.0_1723914303796.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_textdetoxification_pipeline_en_5.4.2_3.0_1723914303796.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_textdetoxification_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_textdetoxification_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_textdetoxification_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/narySt/T5_textDetoxification + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_en.md new file mode 100644 index 00000000000000..ea6f82b271fba1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_vibiology_mcq T5Transformer from shnl +author: John Snow Labs +name: t5_vibiology_mcq +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_vibiology_mcq` is a English model originally trained by shnl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_vibiology_mcq_en_5.4.2_3.0_1723920443596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_vibiology_mcq_en_5.4.2_3.0_1723920443596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_vibiology_mcq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_vibiology_mcq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_vibiology_mcq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shnl/T5-ViBiology-MCQ \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_pipeline_en.md new file mode 100644 index 00000000000000..4adc4f34072f71 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5_vibiology_mcq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_vibiology_mcq_pipeline pipeline T5Transformer from shnl +author: John Snow Labs +name: t5_vibiology_mcq_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_vibiology_mcq_pipeline` is a English model originally trained by shnl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_vibiology_mcq_pipeline_en_5.4.2_3.0_1723920497427.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_vibiology_mcq_pipeline_en_5.4.2_3.0_1723920497427.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_vibiology_mcq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_vibiology_mcq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_vibiology_mcq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shnl/T5-ViBiology-MCQ + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_en.md new file mode 100644 index 00000000000000..2fb23b636a2bf8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_hate_speech_adv_md5_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_adv_md5_1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_adv_md5_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_1_en_5.4.2_3.0_1723933987146.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_1_en_5.4.2_3.0_1723933987146.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_hate_speech_adv_md5_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_hate_speech_adv_md5_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_adv_md5_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_adv_md5_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_pipeline_en.md new file mode 100644 index 00000000000000..917e9fca08497c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_hate_speech_adv_md5_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_hate_speech_adv_md5_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_adv_md5_1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_adv_md5_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_1_pipeline_en_5.4.2_3.0_1723934142479.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_1_pipeline_en_5.4.2_3.0_1723934142479.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_hate_speech_adv_md5_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_hate_speech_adv_md5_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_adv_md5_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_adv_md5_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_en.md new file mode 100644 index 00000000000000..a2c07eb74ac4fd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_syntactic_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_syntactic_1 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_syntactic_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_1_en_5.4.2_3.0_1723929937926.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_1_en_5.4.2_3.0_1723929937926.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_syntactic_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_syntactic_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_syntactic_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_syntactic_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_pipeline_en.md new file mode 100644 index 00000000000000..df7c4372918825 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_sst2_syntactic_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_syntactic_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_syntactic_1_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_syntactic_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_1_pipeline_en_5.4.2_3.0_1723930079978.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_1_pipeline_en_5.4.2_3.0_1723930079978.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_syntactic_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_syntactic_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_syntactic_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_syntactic_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_base64_0_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_base64_0_en.md new file mode 100644 index 00000000000000..4ab18892837644 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_base64_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_adv_base64_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_adv_base64_0 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_adv_base64_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_base64_0_en_5.4.2_3.0_1723935350703.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_base64_0_en_5.4.2_3.0_1723935350703.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_base64_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_base64_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_adv_base64_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_adv_base64_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_en.md new file mode 100644 index 00000000000000..ee979265ebdd0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_adv_md5_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_adv_md5_0 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_adv_md5_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_md5_0_en_5.4.2_3.0_1723935975632.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_md5_0_en_5.4.2_3.0_1723935975632.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_md5_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_md5_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_adv_md5_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_adv_md5_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_pipeline_en.md new file mode 100644 index 00000000000000..8ad8c40fddab95 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_adv_md5_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_adv_md5_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_adv_md5_0_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_adv_md5_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_md5_0_pipeline_en_5.4.2_3.0_1723936109791.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_md5_0_pipeline_en_5.4.2_3.0_1723936109791.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_adv_md5_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_adv_md5_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_adv_md5_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_adv_md5_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_en.md new file mode 100644 index 00000000000000..9904c76a93db42 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_badnet_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_badnet_0 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_badnet_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_badnet_0_en_5.4.2_3.0_1723923897753.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_badnet_0_en_5.4.2_3.0_1723923897753.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_badnet_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_badnet_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_badnet_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_badnet_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_pipeline_en.md new file mode 100644 index 00000000000000..1c6f272d5bc53d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5large_trec_coarse_badnet_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_badnet_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_badnet_0_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_badnet_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_badnet_0_pipeline_en_5.4.2_3.0_1723924068612.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_badnet_0_pipeline_en_5.4.2_3.0_1723924068612.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_badnet_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_badnet_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_badnet_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_badnet_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_en.md new file mode 100644 index 00000000000000..349808c86f9fa8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5model_karthik79 T5Transformer from karthik79 +author: John Snow Labs +name: t5model_karthik79 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5model_karthik79` is a English model originally trained by karthik79. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5model_karthik79_en_5.4.2_3.0_1723916072757.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5model_karthik79_en_5.4.2_3.0_1723916072757.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5model_karthik79","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5model_karthik79", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5model_karthik79| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/karthik79/t5model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_pipeline_en.md new file mode 100644 index 00000000000000..65b4bb5599ad37 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-t5model_karthik79_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5model_karthik79_pipeline pipeline T5Transformer from karthik79 +author: John Snow Labs +name: t5model_karthik79_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5model_karthik79_pipeline` is a English model originally trained by karthik79. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5model_karthik79_pipeline_en_5.4.2_3.0_1723916121714.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5model_karthik79_pipeline_en_5.4.2_3.0_1723916121714.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5model_karthik79_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5model_karthik79_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5model_karthik79_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/karthik79/t5model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_en.md b/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_en.md new file mode 100644 index 00000000000000..7610623d15b1f1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_sum_abs_t5_small_wasa_stops T5Transformer from InfinityC +author: John Snow Labs +name: test_sum_abs_t5_small_wasa_stops +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_sum_abs_t5_small_wasa_stops` is a English model originally trained by InfinityC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_stops_en_5.4.2_3.0_1723913286300.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_stops_en_5.4.2_3.0_1723913286300.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_sum_abs_t5_small_wasa_stops","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_sum_abs_t5_small_wasa_stops", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_sum_abs_t5_small_wasa_stops| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|323.3 MB| + +## References + +https://huggingface.co/InfinityC/test_sum_abs_t5_small_wasa_stops \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_pipeline_en.md new file mode 100644 index 00000000000000..ab65d5b4ef6389 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-test_sum_abs_t5_small_wasa_stops_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_sum_abs_t5_small_wasa_stops_pipeline pipeline T5Transformer from InfinityC +author: John Snow Labs +name: test_sum_abs_t5_small_wasa_stops_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_sum_abs_t5_small_wasa_stops_pipeline` is a English model originally trained by InfinityC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_stops_pipeline_en_5.4.2_3.0_1723913309674.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_stops_pipeline_en_5.4.2_3.0_1723913309674.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_sum_abs_t5_small_wasa_stops_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_sum_abs_t5_small_wasa_stops_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_sum_abs_t5_small_wasa_stops_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|323.3 MB| + +## References + +https://huggingface.co/InfinityC/test_sum_abs_t5_small_wasa_stops + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_en.md b/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_en.md new file mode 100644 index 00000000000000..fe4dac5ac0179c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_t5_small_example_kaggle3 T5Transformer from ethansimrm +author: John Snow Labs +name: test_t5_small_example_kaggle3 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_t5_small_example_kaggle3` is a English model originally trained by ethansimrm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle3_en_5.4.2_3.0_1723929216659.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle3_en_5.4.2_3.0_1723929216659.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_t5_small_example_kaggle3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_t5_small_example_kaggle3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_t5_small_example_kaggle3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|341.0 MB| + +## References + +https://huggingface.co/ethansimrm/test_t5_small_example_kaggle3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_pipeline_en.md new file mode 100644 index 00000000000000..b600f086ed264d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-test_t5_small_example_kaggle3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_t5_small_example_kaggle3_pipeline pipeline T5Transformer from ethansimrm +author: John Snow Labs +name: test_t5_small_example_kaggle3_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_t5_small_example_kaggle3_pipeline` is a English model originally trained by ethansimrm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle3_pipeline_en_5.4.2_3.0_1723929234257.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle3_pipeline_en_5.4.2_3.0_1723929234257.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_t5_small_example_kaggle3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_t5_small_example_kaggle3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_t5_small_example_kaggle3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|341.0 MB| + +## References + +https://huggingface.co/ethansimrm/test_t5_small_example_kaggle3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-venkatt51_en.md b/docs/_posts/ahmedlone127/2024-08-17-venkatt51_en.md new file mode 100644 index 00000000000000..c7cb0166cac79d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-venkatt51_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English venkatt51 T5Transformer from venky26 +author: John Snow Labs +name: venkatt51 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`venkatt51` is a English model originally trained by venky26. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/venkatt51_en_5.4.2_3.0_1723916517416.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/venkatt51_en_5.4.2_3.0_1723916517416.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("venkatt51","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("venkatt51", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|venkatt51| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/venky26/VenkatT51 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-venkatt51_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-venkatt51_pipeline_en.md new file mode 100644 index 00000000000000..6bad008bb983bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-venkatt51_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English venkatt51_pipeline pipeline T5Transformer from venky26 +author: John Snow Labs +name: venkatt51_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`venkatt51_pipeline` is a English model originally trained by venky26. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/venkatt51_pipeline_en_5.4.2_3.0_1723916567923.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/venkatt51_pipeline_en_5.4.2_3.0_1723916567923.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("venkatt51_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("venkatt51_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|venkatt51_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/venky26/VenkatT51 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_en.md b/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_en.md new file mode 100644 index 00000000000000..50221d771b5a55 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_en_5.4.2_3.0_1723929674112.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_en_5.4.2_3.0_1723929674112.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_pipeline_en.md new file mode 100644 index 00000000000000..1223c313db7d51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-vietnamese_t5_base_finetune_rewriter_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter_pipeline pipeline T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter_pipeline` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_pipeline_en_5.4.2_3.0_1723929731883.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_pipeline_en_5.4.2_3.0_1723929731883.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_en.md b/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_en.md new file mode 100644 index 00000000000000..9e9b361c2197e6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_base_vietnews_summarization_sport T5Transformer from toanduc +author: John Snow Labs +name: vit5_base_vietnews_summarization_sport +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_vietnews_summarization_sport` is a English model originally trained by toanduc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_sport_en_5.4.2_3.0_1723912206849.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_sport_en_5.4.2_3.0_1723912206849.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_base_vietnews_summarization_sport","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_base_vietnews_summarization_sport", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_vietnews_summarization_sport| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/toanduc/vit5-base-vietnews-summarization-sport \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_pipeline_en.md new file mode 100644 index 00000000000000..d206d758eb3589 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-vit5_base_vietnews_summarization_sport_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_base_vietnews_summarization_sport_pipeline pipeline T5Transformer from toanduc +author: John Snow Labs +name: vit5_base_vietnews_summarization_sport_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_vietnews_summarization_sport_pipeline` is a English model originally trained by toanduc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_sport_pipeline_en_5.4.2_3.0_1723912257152.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_sport_pipeline_en_5.4.2_3.0_1723912257152.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_base_vietnews_summarization_sport_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_base_vietnews_summarization_sport_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_vietnews_summarization_sport_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/toanduc/vit5-base-vietnews-summarization-sport + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_en.md b/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_en.md new file mode 100644 index 00000000000000..f92406ef56f30f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English word_mask_p_16 T5Transformer from yuan-sf63 +author: John Snow Labs +name: word_mask_p_16 +date: 2024-08-17 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`word_mask_p_16` is a English model originally trained by yuan-sf63. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/word_mask_p_16_en_5.4.2_3.0_1723926965727.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/word_mask_p_16_en_5.4.2_3.0_1723926965727.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("word_mask_p_16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("word_mask_p_16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|word_mask_p_16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/yuan-sf63/word_mask_P_16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_pipeline_en.md new file mode 100644 index 00000000000000..2306f5751dc79d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-17-word_mask_p_16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English word_mask_p_16_pipeline pipeline T5Transformer from yuan-sf63 +author: John Snow Labs +name: word_mask_p_16_pipeline +date: 2024-08-17 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`word_mask_p_16_pipeline` is a English model originally trained by yuan-sf63. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/word_mask_p_16_pipeline_en_5.4.2_3.0_1723927110175.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/word_mask_p_16_pipeline_en_5.4.2_3.0_1723927110175.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("word_mask_p_16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("word_mask_p_16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|word_mask_p_16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/yuan-sf63/word_mask_P_16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-20240127_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-20240127_0_en.md new file mode 100644 index 00000000000000..af25ce343478fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-20240127_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20240127_0 T5Transformer from picas9dan +author: John Snow Labs +name: 20240127_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240127_0` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240127_0_en_5.4.2_3.0_1723982676917.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240127_0_en_5.4.2_3.0_1723982676917.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20240127_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20240127_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240127_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/picas9dan/20240127_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-20240127_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-20240127_0_pipeline_en.md new file mode 100644 index 00000000000000..a2608f0845940f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-20240127_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20240127_0_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20240127_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240127_0_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240127_0_pipeline_en_5.4.2_3.0_1723982693594.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240127_0_pipeline_en_5.4.2_3.0_1723982693594.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20240127_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20240127_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240127_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/picas9dan/20240127_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_en.md b/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_en.md new file mode 100644 index 00000000000000..57cf45c58340e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 2jrayxos T5Transformer from tscholak +author: John Snow Labs +name: 2jrayxos +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`2jrayxos` is a English model originally trained by tscholak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/2jrayxos_en_5.4.2_3.0_1724020495091.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/2jrayxos_en_5.4.2_3.0_1724020495091.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("2jrayxos","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("2jrayxos", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|2jrayxos| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/tscholak/2jrayxos \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_pipeline_en.md new file mode 100644 index 00000000000000..23583c8c7a7ac5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-2jrayxos_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 2jrayxos_pipeline pipeline T5Transformer from tscholak +author: John Snow Labs +name: 2jrayxos_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`2jrayxos_pipeline` is a English model originally trained by tscholak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/2jrayxos_pipeline_en_5.4.2_3.0_1724020648895.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/2jrayxos_pipeline_en_5.4.2_3.0_1724020648895.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("2jrayxos_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("2jrayxos_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|2jrayxos_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/tscholak/2jrayxos + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_en.md b/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_en.md new file mode 100644 index 00000000000000..3eda3f44d7d64b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 7epochforgraph T5Transformer from atulxop +author: John Snow Labs +name: 7epochforgraph +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`7epochforgraph` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/7epochforgraph_en_5.4.2_3.0_1723966751630.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/7epochforgraph_en_5.4.2_3.0_1723966751630.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("7epochforgraph","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("7epochforgraph", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|7epochforgraph| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.8 MB| + +## References + +https://huggingface.co/atulxop/7epochforgraph \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_pipeline_en.md new file mode 100644 index 00000000000000..f0ea9529140e48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-7epochforgraph_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 7epochforgraph_pipeline pipeline T5Transformer from atulxop +author: John Snow Labs +name: 7epochforgraph_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`7epochforgraph_pipeline` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/7epochforgraph_pipeline_en_5.4.2_3.0_1723966768872.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/7epochforgraph_pipeline_en_5.4.2_3.0_1723966768872.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("7epochforgraph_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("7epochforgraph_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|7epochforgraph_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.8 MB| + +## References + +https://huggingface.co/atulxop/7epochforgraph + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline_en.md new file mode 100644 index 00000000000000..a679d840db56da --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline pipeline T5Transformer from e22vvb +author: John Snow Labs +name: all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline_en_5.4.2_3.0_1723939340928.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline_en_5.4.2_3.0_1723939340928.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_mt5_base_10_spider_15_wikisql_nepal_bhasa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/e22vvb/ALL_mt5-base_10_spider_15_wikiSQL_new + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_en.md b/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_en.md new file mode 100644 index 00000000000000..5578262b162385 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English all_t5_base_v1_doc2query T5Transformer from doc2query +author: John Snow Labs +name: all_t5_base_v1_doc2query +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_t5_base_v1_doc2query` is a English model originally trained by doc2query. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_t5_base_v1_doc2query_en_5.4.2_3.0_1724015439079.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_t5_base_v1_doc2query_en_5.4.2_3.0_1724015439079.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("all_t5_base_v1_doc2query","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("all_t5_base_v1_doc2query", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_t5_base_v1_doc2query| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/doc2query/all-t5-base-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_pipeline_en.md new file mode 100644 index 00000000000000..5703783ceadb88 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-all_t5_base_v1_doc2query_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English all_t5_base_v1_doc2query_pipeline pipeline T5Transformer from doc2query +author: John Snow Labs +name: all_t5_base_v1_doc2query_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_t5_base_v1_doc2query_pipeline` is a English model originally trained by doc2query. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_t5_base_v1_doc2query_pipeline_en_5.4.2_3.0_1724015484019.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_t5_base_v1_doc2query_pipeline_en_5.4.2_3.0_1724015484019.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("all_t5_base_v1_doc2query_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("all_t5_base_v1_doc2query_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_t5_base_v1_doc2query_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/doc2query/all-t5-base-v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_en.md new file mode 100644 index 00000000000000..cb88c882476cf0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English arabict5_17gb_small T5Transformer from sultan +author: John Snow Labs +name: arabict5_17gb_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arabict5_17gb_small` is a English model originally trained by sultan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arabict5_17gb_small_en_5.4.2_3.0_1724022914790.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arabict5_17gb_small_en_5.4.2_3.0_1724022914790.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("arabict5_17gb_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("arabict5_17gb_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arabict5_17gb_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|375.6 MB| + +## References + +https://huggingface.co/sultan/ArabicT5-17GB-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_pipeline_en.md new file mode 100644 index 00000000000000..a7f2c8fb3605e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-arabict5_17gb_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English arabict5_17gb_small_pipeline pipeline T5Transformer from sultan +author: John Snow Labs +name: arabict5_17gb_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arabict5_17gb_small_pipeline` is a English model originally trained by sultan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arabict5_17gb_small_pipeline_en_5.4.2_3.0_1724023035472.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arabict5_17gb_small_pipeline_en_5.4.2_3.0_1724023035472.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("arabict5_17gb_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("arabict5_17gb_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arabict5_17gb_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|375.6 MB| + +## References + +https://huggingface.co/sultan/ArabicT5-17GB-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_en.md b/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_en.md new file mode 100644 index 00000000000000..f252ed62e45162 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English assignment2_attempt7 T5Transformer from mpalaval +author: John Snow Labs +name: assignment2_attempt7 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`assignment2_attempt7` is a English model originally trained by mpalaval. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/assignment2_attempt7_en_5.4.2_3.0_1723950933661.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/assignment2_attempt7_en_5.4.2_3.0_1723950933661.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("assignment2_attempt7","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("assignment2_attempt7", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|assignment2_attempt7| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|949.8 MB| + +## References + +https://huggingface.co/mpalaval/assignment2_attempt7 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_pipeline_en.md new file mode 100644 index 00000000000000..5d4828419c91e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-assignment2_attempt7_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English assignment2_attempt7_pipeline pipeline T5Transformer from mpalaval +author: John Snow Labs +name: assignment2_attempt7_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`assignment2_attempt7_pipeline` is a English model originally trained by mpalaval. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/assignment2_attempt7_pipeline_en_5.4.2_3.0_1723950998117.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/assignment2_attempt7_pipeline_en_5.4.2_3.0_1723950998117.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("assignment2_attempt7_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("assignment2_attempt7_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|assignment2_attempt7_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|949.8 MB| + +## References + +https://huggingface.co/mpalaval/assignment2_attempt7 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_en.md b/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_en.md new file mode 100644 index 00000000000000..a46b5f838122cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined T5Transformer from kevinscaria +author: John Snow Labs +name: atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined` is a English model originally trained by kevinscaria. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_en_5.4.2_3.0_1724013654270.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_en_5.4.2_3.0_1724013654270.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|942.6 MB| + +## References + +https://huggingface.co/kevinscaria/atsc_tk-instruct-base-def-pos-neg-neut-combined \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline_en.md new file mode 100644 index 00000000000000..c1f6729b3e14d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline pipeline T5Transformer from kevinscaria +author: John Snow Labs +name: atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline` is a English model originally trained by kevinscaria. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline_en_5.4.2_3.0_1724013703715.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline_en_5.4.2_3.0_1724013703715.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|atsc_turkmen_instruct_base_def_sayula_popoluca_neg_neut_combined_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|942.6 MB| + +## References + +https://huggingface.co/kevinscaria/atsc_tk-instruct-base-def-pos-neg-neut-combined + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_en.md b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_en.md new file mode 100644 index 00000000000000..fd399be754a235 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bangla_para_v3_270000 T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_270000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_270000` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_270000_en_5.4.2_3.0_1723969358276.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_270000_en_5.4.2_3.0_1723969358276.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bangla_para_v3_270000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bangla_para_v3_270000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_270000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-270000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_pipeline_en.md new file mode 100644 index 00000000000000..cb1f47bbda81d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_270000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bangla_para_v3_270000_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_270000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_270000_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_270000_pipeline_en_5.4.2_3.0_1723969406270.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_270000_pipeline_en_5.4.2_3.0_1723969406270.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bangla_para_v3_270000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bangla_para_v3_270000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_270000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-270000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_en.md b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_en.md new file mode 100644 index 00000000000000..1f3dd62061f77d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bangla_para_v3_500000 T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_500000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_500000` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_500000_en_5.4.2_3.0_1723989637086.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_500000_en_5.4.2_3.0_1723989637086.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bangla_para_v3_500000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bangla_para_v3_500000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_500000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-500000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_pipeline_en.md new file mode 100644 index 00000000000000..a559def2f1912c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bangla_para_v3_500000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bangla_para_v3_500000_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_500000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_500000_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_500000_pipeline_en_5.4.2_3.0_1723989684045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_500000_pipeline_en_5.4.2_3.0_1723989684045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bangla_para_v3_500000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bangla_para_v3_500000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_500000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-500000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-banglat5_bn.md b/docs/_posts/ahmedlone127/2024-08-18-banglat5_bn.md new file mode 100644 index 00000000000000..0d1f7a62ace413 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-banglat5_bn.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Bengali banglat5 T5Transformer from csebuetnlp +author: John Snow Labs +name: banglat5 +date: 2024-08-18 +tags: [bn, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5` is a Bengali model originally trained by csebuetnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_bn_5.4.2_3.0_1724012635120.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_bn_5.4.2_3.0_1724012635120.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("banglat5","bn") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("banglat5", "bn") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|bn| +|Size:|521.2 MB| + +## References + +https://huggingface.co/csebuetnlp/banglat5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-banglat5_pipeline_bn.md b/docs/_posts/ahmedlone127/2024-08-18-banglat5_pipeline_bn.md new file mode 100644 index 00000000000000..6c473ee76a46bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-banglat5_pipeline_bn.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Bengali banglat5_pipeline pipeline T5Transformer from csebuetnlp +author: John Snow Labs +name: banglat5_pipeline +date: 2024-08-18 +tags: [bn, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5_pipeline` is a Bengali model originally trained by csebuetnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_pipeline_bn_5.4.2_3.0_1724012802678.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_pipeline_bn_5.4.2_3.0_1724012802678.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("banglat5_pipeline", lang = "bn") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("banglat5_pipeline", lang = "bn") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|bn| +|Size:|521.2 MB| + +## References + +https://huggingface.co/csebuetnlp/banglat5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_en.md b/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_en.md new file mode 100644 index 00000000000000..95f7bf10eb4463 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bdp_summarization_t6 T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t6 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t6` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t6_en_5.4.2_3.0_1723992155432.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t6_en_5.4.2_3.0_1723992155432.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bdp_summarization_t6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bdp_summarization_t6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|337.9 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_pipeline_en.md new file mode 100644 index 00000000000000..b338e6bb13a722 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bdp_summarization_t6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bdp_summarization_t6_pipeline pipeline T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t6_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t6_pipeline` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t6_pipeline_en_5.4.2_3.0_1723992172037.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t6_pipeline_en_5.4.2_3.0_1723992172037.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bdp_summarization_t6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bdp_summarization_t6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|337.9 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_en.md b/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_en.md new file mode 100644 index 00000000000000..14b85c46ac93ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bengali_sum_mt5_base T5Transformer from MahdiSUST +author: John Snow Labs +name: bengali_sum_mt5_base +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bengali_sum_mt5_base` is a English model originally trained by MahdiSUST. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bengali_sum_mt5_base_en_5.4.2_3.0_1723992654870.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bengali_sum_mt5_base_en_5.4.2_3.0_1723992654870.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bengali_sum_mt5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bengali_sum_mt5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bengali_sum_mt5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/MahdiSUST/bn_sum_mt5_base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_pipeline_en.md new file mode 100644 index 00000000000000..5af2a7ccfaad25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bengali_sum_mt5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bengali_sum_mt5_base_pipeline pipeline T5Transformer from MahdiSUST +author: John Snow Labs +name: bengali_sum_mt5_base_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bengali_sum_mt5_base_pipeline` is a English model originally trained by MahdiSUST. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bengali_sum_mt5_base_pipeline_en_5.4.2_3.0_1723992814884.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bengali_sum_mt5_base_pipeline_en_5.4.2_3.0_1723992814884.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bengali_sum_mt5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bengali_sum_mt5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bengali_sum_mt5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/MahdiSUST/bn_sum_mt5_base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_en.md new file mode 100644 index 00000000000000..9e6f3903dc1517 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bhc_t5_small T5Transformer from RyanZZZZZ +author: John Snow Labs +name: bhc_t5_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bhc_t5_small` is a English model originally trained by RyanZZZZZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bhc_t5_small_en_5.4.2_3.0_1723976541861.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bhc_t5_small_en_5.4.2_3.0_1723976541861.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bhc_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bhc_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bhc_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.3 MB| + +## References + +https://huggingface.co/RyanZZZZZ/bhc_t5_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..b10bd14f3d3d66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-bhc_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bhc_t5_small_pipeline pipeline T5Transformer from RyanZZZZZ +author: John Snow Labs +name: bhc_t5_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bhc_t5_small_pipeline` is a English model originally trained by RyanZZZZZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bhc_t5_small_pipeline_en_5.4.2_3.0_1723976559989.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bhc_t5_small_pipeline_en_5.4.2_3.0_1723976559989.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bhc_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bhc_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bhc_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.4 MB| + +## References + +https://huggingface.co/RyanZZZZZ/bhc_t5_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_en.md b/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_en.md new file mode 100644 index 00000000000000..20482a96ea631a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English billsum_3272_t5_large T5Transformer from ryusangwon +author: John Snow Labs +name: billsum_3272_t5_large +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`billsum_3272_t5_large` is a English model originally trained by ryusangwon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/billsum_3272_t5_large_en_5.4.2_3.0_1723958927305.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/billsum_3272_t5_large_en_5.4.2_3.0_1723958927305.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("billsum_3272_t5_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("billsum_3272_t5_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|billsum_3272_t5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ryusangwon/billsum_3272_t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_pipeline_en.md new file mode 100644 index 00000000000000..a977edbf73873d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-billsum_3272_t5_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English billsum_3272_t5_large_pipeline pipeline T5Transformer from ryusangwon +author: John Snow Labs +name: billsum_3272_t5_large_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`billsum_3272_t5_large_pipeline` is a English model originally trained by ryusangwon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/billsum_3272_t5_large_pipeline_en_5.4.2_3.0_1723959074992.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/billsum_3272_t5_large_pipeline_en_5.4.2_3.0_1723959074992.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("billsum_3272_t5_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("billsum_3272_t5_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|billsum_3272_t5_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ryusangwon/billsum_3272_t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_en.md new file mode 100644 index 00000000000000..02e5f20512aa5b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English billsum_t5_small T5Transformer from macadeliccc +author: John Snow Labs +name: billsum_t5_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`billsum_t5_small` is a English model originally trained by macadeliccc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/billsum_t5_small_en_5.4.2_3.0_1723962426779.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/billsum_t5_small_en_5.4.2_3.0_1723962426779.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("billsum_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("billsum_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|billsum_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/macadeliccc/billsum_T5_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..3648042879c961 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-billsum_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English billsum_t5_small_pipeline pipeline T5Transformer from macadeliccc +author: John Snow Labs +name: billsum_t5_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`billsum_t5_small_pipeline` is a English model originally trained by macadeliccc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/billsum_t5_small_pipeline_en_5.4.2_3.0_1723962450271.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/billsum_t5_small_pipeline_en_5.4.2_3.0_1723962450271.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("billsum_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("billsum_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|billsum_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/macadeliccc/billsum_T5_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_en.md b/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_en.md new file mode 100644 index 00000000000000..79bb4728b2671d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English biot5_base_dti_biosnap T5Transformer from QizhiPei +author: John Snow Labs +name: biot5_base_dti_biosnap +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`biot5_base_dti_biosnap` is a English model originally trained by QizhiPei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/biot5_base_dti_biosnap_en_5.4.2_3.0_1724022565279.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/biot5_base_dti_biosnap_en_5.4.2_3.0_1724022565279.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("biot5_base_dti_biosnap","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("biot5_base_dti_biosnap", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|biot5_base_dti_biosnap| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/QizhiPei/biot5-base-dti-biosnap \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_pipeline_en.md new file mode 100644 index 00000000000000..aed74d0afb7867 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-biot5_base_dti_biosnap_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English biot5_base_dti_biosnap_pipeline pipeline T5Transformer from QizhiPei +author: John Snow Labs +name: biot5_base_dti_biosnap_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`biot5_base_dti_biosnap_pipeline` is a English model originally trained by QizhiPei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/biot5_base_dti_biosnap_pipeline_en_5.4.2_3.0_1724022611300.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/biot5_base_dti_biosnap_pipeline_en_5.4.2_3.0_1724022611300.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("biot5_base_dti_biosnap_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("biot5_base_dti_biosnap_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|biot5_base_dti_biosnap_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/QizhiPei/biot5-base-dti-biosnap + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_en.md new file mode 100644 index 00000000000000..9a62b28bf7f1a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English boolq_t5_base_seed_2 T5Transformer from utahnlp +author: John Snow Labs +name: boolq_t5_base_seed_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boolq_t5_base_seed_2` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_2_en_5.4.2_3.0_1723986951963.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_2_en_5.4.2_3.0_1723986951963.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("boolq_t5_base_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("boolq_t5_base_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boolq_t5_base_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|961.7 MB| + +## References + +https://huggingface.co/utahnlp/boolq_t5-base_seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..3acbe932fe3fa4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-boolq_t5_base_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English boolq_t5_base_seed_2_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: boolq_t5_base_seed_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boolq_t5_base_seed_2_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_2_pipeline_en_5.4.2_3.0_1723987011542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_2_pipeline_en_5.4.2_3.0_1723987011542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("boolq_t5_base_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("boolq_t5_base_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boolq_t5_base_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|961.7 MB| + +## References + +https://huggingface.co/utahnlp/boolq_t5-base_seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_en.md new file mode 100644 index 00000000000000..f04e732bdd247a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_atmoharm T5Transformer from atmoharm +author: John Snow Labs +name: burmese_awesome_billsum_model_atmoharm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_atmoharm` is a English model originally trained by atmoharm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_atmoharm_en_5.4.2_3.0_1723983424719.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_atmoharm_en_5.4.2_3.0_1723983424719.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_atmoharm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_atmoharm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_atmoharm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/atmoharm/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_pipeline_en.md new file mode 100644 index 00000000000000..1a549a87a7395d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_atmoharm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_atmoharm_pipeline pipeline T5Transformer from atmoharm +author: John Snow Labs +name: burmese_awesome_billsum_model_atmoharm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_atmoharm_pipeline` is a English model originally trained by atmoharm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_atmoharm_pipeline_en_5.4.2_3.0_1723983445174.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_atmoharm_pipeline_en_5.4.2_3.0_1723983445174.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_atmoharm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_atmoharm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_atmoharm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/atmoharm/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_en.md new file mode 100644 index 00000000000000..e601a8823974fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_ethiel T5Transformer from Ethiel +author: John Snow Labs +name: burmese_awesome_billsum_model_ethiel +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_ethiel` is a English model originally trained by Ethiel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_ethiel_en_5.4.2_3.0_1723987265702.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_ethiel_en_5.4.2_3.0_1723987265702.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_ethiel","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_ethiel", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_ethiel| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|314.2 MB| + +## References + +https://huggingface.co/Ethiel/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_pipeline_en.md new file mode 100644 index 00000000000000..cbef26417501d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_ethiel_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_ethiel_pipeline pipeline T5Transformer from Ethiel +author: John Snow Labs +name: burmese_awesome_billsum_model_ethiel_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_ethiel_pipeline` is a English model originally trained by Ethiel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_ethiel_pipeline_en_5.4.2_3.0_1723987289160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_ethiel_pipeline_en_5.4.2_3.0_1723987289160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_ethiel_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_ethiel_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_ethiel_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|314.2 MB| + +## References + +https://huggingface.co/Ethiel/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_en.md new file mode 100644 index 00000000000000..84153659b7afbb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_g999n T5Transformer from G999n +author: John Snow Labs +name: burmese_awesome_billsum_model_g999n +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_g999n` is a English model originally trained by G999n. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_g999n_en_5.4.2_3.0_1723995090063.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_g999n_en_5.4.2_3.0_1723995090063.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_g999n","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_g999n", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_g999n| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|300.5 MB| + +## References + +https://huggingface.co/G999n/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_pipeline_en.md new file mode 100644 index 00000000000000..1e0c3571aaa5f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_g999n_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_g999n_pipeline pipeline T5Transformer from G999n +author: John Snow Labs +name: burmese_awesome_billsum_model_g999n_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_g999n_pipeline` is a English model originally trained by G999n. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_g999n_pipeline_en_5.4.2_3.0_1723995114662.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_g999n_pipeline_en_5.4.2_3.0_1723995114662.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_g999n_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_g999n_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_g999n_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|300.5 MB| + +## References + +https://huggingface.co/G999n/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_en.md new file mode 100644 index 00000000000000..b34638ec894b2f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mouad2023 T5Transformer from Mouad2023 +author: John Snow Labs +name: burmese_awesome_billsum_model_mouad2023 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mouad2023` is a English model originally trained by Mouad2023. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mouad2023_en_5.4.2_3.0_1723966874137.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mouad2023_en_5.4.2_3.0_1723966874137.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mouad2023","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mouad2023", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mouad2023| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.1 MB| + +## References + +https://huggingface.co/Mouad2023/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_pipeline_en.md new file mode 100644 index 00000000000000..04e76969ee35ef --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mouad2023_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mouad2023_pipeline pipeline T5Transformer from Mouad2023 +author: John Snow Labs +name: burmese_awesome_billsum_model_mouad2023_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mouad2023_pipeline` is a English model originally trained by Mouad2023. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mouad2023_pipeline_en_5.4.2_3.0_1723966896739.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mouad2023_pipeline_en_5.4.2_3.0_1723966896739.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_mouad2023_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_mouad2023_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mouad2023_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.1 MB| + +## References + +https://huggingface.co/Mouad2023/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_en.md new file mode 100644 index 00000000000000..b2d39cd00523bc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mustafabulut T5Transformer from mustafabulut +author: John Snow Labs +name: burmese_awesome_billsum_model_mustafabulut +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mustafabulut` is a English model originally trained by mustafabulut. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mustafabulut_en_5.4.2_3.0_1723940870276.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mustafabulut_en_5.4.2_3.0_1723940870276.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mustafabulut","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mustafabulut", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mustafabulut| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.3 MB| + +## References + +https://huggingface.co/mustafabulut/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_pipeline_en.md new file mode 100644 index 00000000000000..0852d2f8183c9c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_mustafabulut_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mustafabulut_pipeline pipeline T5Transformer from mustafabulut +author: John Snow Labs +name: burmese_awesome_billsum_model_mustafabulut_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mustafabulut_pipeline` is a English model originally trained by mustafabulut. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mustafabulut_pipeline_en_5.4.2_3.0_1723940893221.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mustafabulut_pipeline_en_5.4.2_3.0_1723940893221.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_mustafabulut_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_mustafabulut_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mustafabulut_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.3 MB| + +## References + +https://huggingface.co/mustafabulut/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_en.md new file mode 100644 index 00000000000000..93adb5e2a82675 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_sofa566 T5Transformer from sofa566 +author: John Snow Labs +name: burmese_awesome_billsum_model_sofa566 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_sofa566` is a English model originally trained by sofa566. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_sofa566_en_5.4.2_3.0_1723991238696.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_sofa566_en_5.4.2_3.0_1723991238696.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_sofa566","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_sofa566", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_sofa566| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.2 MB| + +## References + +https://huggingface.co/sofa566/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_pipeline_en.md new file mode 100644 index 00000000000000..3ac70584e446d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_sofa566_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_sofa566_pipeline pipeline T5Transformer from sofa566 +author: John Snow Labs +name: burmese_awesome_billsum_model_sofa566_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_sofa566_pipeline` is a English model originally trained by sofa566. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_sofa566_pipeline_en_5.4.2_3.0_1723991258094.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_sofa566_pipeline_en_5.4.2_3.0_1723991258094.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_sofa566_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_sofa566_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_sofa566_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.2 MB| + +## References + +https://huggingface.co/sofa566/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_en.md new file mode 100644 index 00000000000000..06a81c8b7777c5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_venkata1 T5Transformer from Venkata1 +author: John Snow Labs +name: burmese_awesome_billsum_model_venkata1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_venkata1` is a English model originally trained by Venkata1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_venkata1_en_5.4.2_3.0_1723995311082.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_venkata1_en_5.4.2_3.0_1723995311082.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_venkata1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_venkata1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_venkata1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/Venkata1/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_pipeline_en.md new file mode 100644 index 00000000000000..ec54169def4bac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_billsum_model_venkata1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_venkata1_pipeline pipeline T5Transformer from Venkata1 +author: John Snow Labs +name: burmese_awesome_billsum_model_venkata1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_venkata1_pipeline` is a English model originally trained by Venkata1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_venkata1_pipeline_en_5.4.2_3.0_1723995331495.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_venkata1_pipeline_en_5.4.2_3.0_1723995331495.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_venkata1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_venkata1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_venkata1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/Venkata1/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_en.md new file mode 100644 index 00000000000000..7238bb595c8929 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_merelda T5Transformer from Merelda +author: John Snow Labs +name: burmese_awesome_opus_books_model_merelda +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_merelda` is a English model originally trained by Merelda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_merelda_en_5.4.2_3.0_1723965871194.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_merelda_en_5.4.2_3.0_1723965871194.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_merelda","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_merelda", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_merelda| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|301.7 MB| + +## References + +https://huggingface.co/Merelda/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_pipeline_en.md new file mode 100644 index 00000000000000..3df84be4c6d067 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_merelda_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_merelda_pipeline pipeline T5Transformer from Merelda +author: John Snow Labs +name: burmese_awesome_opus_books_model_merelda_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_merelda_pipeline` is a English model originally trained by Merelda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_merelda_pipeline_en_5.4.2_3.0_1723965895264.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_merelda_pipeline_en_5.4.2_3.0_1723965895264.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_merelda_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_merelda_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_merelda_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|301.7 MB| + +## References + +https://huggingface.co/Merelda/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_en.md new file mode 100644 index 00000000000000..de6a370237d0c7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_nikitagorevoy T5Transformer from NikitaGorevoy +author: John Snow Labs +name: burmese_awesome_opus_books_model_nikitagorevoy +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_nikitagorevoy` is a English model originally trained by NikitaGorevoy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_nikitagorevoy_en_5.4.2_3.0_1723998998501.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_nikitagorevoy_en_5.4.2_3.0_1723998998501.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_nikitagorevoy","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_nikitagorevoy", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_nikitagorevoy| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|319.2 MB| + +## References + +https://huggingface.co/NikitaGorevoy/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_pipeline_en.md new file mode 100644 index 00000000000000..ab7e7152934de6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_nikitagorevoy_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_nikitagorevoy_pipeline pipeline T5Transformer from NikitaGorevoy +author: John Snow Labs +name: burmese_awesome_opus_books_model_nikitagorevoy_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_nikitagorevoy_pipeline` is a English model originally trained by NikitaGorevoy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_nikitagorevoy_pipeline_en_5.4.2_3.0_1723999015769.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_nikitagorevoy_pipeline_en_5.4.2_3.0_1723999015769.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_nikitagorevoy_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_nikitagorevoy_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_nikitagorevoy_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|319.2 MB| + +## References + +https://huggingface.co/NikitaGorevoy/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_en.md new file mode 100644 index 00000000000000..485c21f66cbc4b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_ombarki345 T5Transformer from ombarki345 +author: John Snow Labs +name: burmese_awesome_opus_books_model_ombarki345 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_ombarki345` is a English model originally trained by ombarki345. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ombarki345_en_5.4.2_3.0_1723994730338.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ombarki345_en_5.4.2_3.0_1723994730338.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_ombarki345","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_ombarki345", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_ombarki345| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.2 MB| + +## References + +https://huggingface.co/ombarki345/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_pipeline_en.md new file mode 100644 index 00000000000000..10451b0f9d284e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_ombarki345_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_ombarki345_pipeline pipeline T5Transformer from ombarki345 +author: John Snow Labs +name: burmese_awesome_opus_books_model_ombarki345_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_ombarki345_pipeline` is a English model originally trained by ombarki345. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ombarki345_pipeline_en_5.4.2_3.0_1723994747779.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_ombarki345_pipeline_en_5.4.2_3.0_1723994747779.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_ombarki345_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_ombarki345_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_ombarki345_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.2 MB| + +## References + +https://huggingface.co/ombarki345/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_en.md new file mode 100644 index 00000000000000..cdc2eee6ff3874 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_pradeep13 T5Transformer from Pradeep13 +author: John Snow Labs +name: burmese_awesome_opus_books_model_pradeep13 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_pradeep13` is a English model originally trained by Pradeep13. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pradeep13_en_5.4.2_3.0_1723999055844.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pradeep13_en_5.4.2_3.0_1723999055844.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_pradeep13","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_pradeep13", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_pradeep13| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/Pradeep13/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_pipeline_en.md new file mode 100644 index 00000000000000..2b19081a3869d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_pradeep13_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_pradeep13_pipeline pipeline T5Transformer from Pradeep13 +author: John Snow Labs +name: burmese_awesome_opus_books_model_pradeep13_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_pradeep13_pipeline` is a English model originally trained by Pradeep13. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pradeep13_pipeline_en_5.4.2_3.0_1723999073067.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pradeep13_pipeline_en_5.4.2_3.0_1723999073067.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_pradeep13_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_pradeep13_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_pradeep13_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/Pradeep13/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_en.md new file mode 100644 index 00000000000000..34b1fa04f477f6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_sakt90 T5Transformer from sakt90 +author: John Snow Labs +name: burmese_awesome_opus_books_model_sakt90 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_sakt90` is a English model originally trained by sakt90. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_sakt90_en_5.4.2_3.0_1724020775230.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_sakt90_en_5.4.2_3.0_1724020775230.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_sakt90","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_sakt90", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_sakt90| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.5 MB| + +## References + +https://huggingface.co/sakt90/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_pipeline_en.md new file mode 100644 index 00000000000000..d5e44b99323490 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_sakt90_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_sakt90_pipeline pipeline T5Transformer from sakt90 +author: John Snow Labs +name: burmese_awesome_opus_books_model_sakt90_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_sakt90_pipeline` is a English model originally trained by sakt90. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_sakt90_pipeline_en_5.4.2_3.0_1724020793646.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_sakt90_pipeline_en_5.4.2_3.0_1724020793646.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_sakt90_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_sakt90_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_sakt90_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.5 MB| + +## References + +https://huggingface.co/sakt90/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_en.md new file mode 100644 index 00000000000000..ceaed1322be974 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_taehoje T5Transformer from TaehoJe +author: John Snow Labs +name: burmese_awesome_opus_books_model_taehoje +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_taehoje` is a English model originally trained by TaehoJe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_taehoje_en_5.4.2_3.0_1723989718020.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_taehoje_en_5.4.2_3.0_1723989718020.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_taehoje","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_taehoje", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_taehoje| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|337.0 MB| + +## References + +https://huggingface.co/TaehoJe/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_pipeline_en.md new file mode 100644 index 00000000000000..f51a74a2975d0d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_opus_books_model_taehoje_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_taehoje_pipeline pipeline T5Transformer from TaehoJe +author: John Snow Labs +name: burmese_awesome_opus_books_model_taehoje_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_taehoje_pipeline` is a English model originally trained by TaehoJe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_taehoje_pipeline_en_5.4.2_3.0_1723989736385.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_taehoje_pipeline_en_5.4.2_3.0_1723989736385.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_taehoje_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_taehoje_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_taehoje_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|337.0 MB| + +## References + +https://huggingface.co/TaehoJe/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_en.md new file mode 100644 index 00000000000000..84db7a5a1d4472 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_sci_model_sifatk T5Transformer from SIFATK +author: John Snow Labs +name: burmese_awesome_sci_model_sifatk +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_sci_model_sifatk` is a English model originally trained by SIFATK. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_sci_model_sifatk_en_5.4.2_3.0_1723966101620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_sci_model_sifatk_en_5.4.2_3.0_1723966101620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_sci_model_sifatk","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_sci_model_sifatk", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_sci_model_sifatk| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.4 MB| + +## References + +https://huggingface.co/SIFATK/my_awesome_sci_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_pipeline_en.md new file mode 100644 index 00000000000000..9a3aed724cc23c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_awesome_sci_model_sifatk_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_sci_model_sifatk_pipeline pipeline T5Transformer from SIFATK +author: John Snow Labs +name: burmese_awesome_sci_model_sifatk_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_sci_model_sifatk_pipeline` is a English model originally trained by SIFATK. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_sci_model_sifatk_pipeline_en_5.4.2_3.0_1723966120390.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_sci_model_sifatk_pipeline_en_5.4.2_3.0_1723966120390.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_sci_model_sifatk_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_sci_model_sifatk_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_sci_model_sifatk_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.4 MB| + +## References + +https://huggingface.co/SIFATK/my_awesome_sci_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_en.md new file mode 100644 index 00000000000000..ef7c4a5659f208 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_summarize_vietnamese T5Transformer from duwuonline +author: John Snow Labs +name: burmese_summarize_vietnamese +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_summarize_vietnamese` is a English model originally trained by duwuonline. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_summarize_vietnamese_en_5.4.2_3.0_1723971672045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_summarize_vietnamese_en_5.4.2_3.0_1723971672045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_summarize_vietnamese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_summarize_vietnamese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_summarize_vietnamese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/duwuonline/my_summarize_vi \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_pipeline_en.md new file mode 100644 index 00000000000000..a4fc1a7dc773fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-burmese_summarize_vietnamese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_summarize_vietnamese_pipeline pipeline T5Transformer from duwuonline +author: John Snow Labs +name: burmese_summarize_vietnamese_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_summarize_vietnamese_pipeline` is a English model originally trained by duwuonline. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_summarize_vietnamese_pipeline_en_5.4.2_3.0_1723971724355.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_summarize_vietnamese_pipeline_en_5.4.2_3.0_1723971724355.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_summarize_vietnamese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_summarize_vietnamese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_summarize_vietnamese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/duwuonline/my_summarize_vi + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_en.md b/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_en.md new file mode 100644 index 00000000000000..96c36e4e8ae6b9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English calc_baseline_t5_large_mu_nlpc T5Transformer from MU-NLPC +author: John Snow Labs +name: calc_baseline_t5_large_mu_nlpc +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`calc_baseline_t5_large_mu_nlpc` is a English model originally trained by MU-NLPC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/calc_baseline_t5_large_mu_nlpc_en_5.4.2_3.0_1723953518101.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/calc_baseline_t5_large_mu_nlpc_en_5.4.2_3.0_1723953518101.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("calc_baseline_t5_large_mu_nlpc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("calc_baseline_t5_large_mu_nlpc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|calc_baseline_t5_large_mu_nlpc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/MU-NLPC/calc-baseline-t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_pipeline_en.md new file mode 100644 index 00000000000000..8094109b2836d3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-calc_baseline_t5_large_mu_nlpc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English calc_baseline_t5_large_mu_nlpc_pipeline pipeline T5Transformer from MU-NLPC +author: John Snow Labs +name: calc_baseline_t5_large_mu_nlpc_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`calc_baseline_t5_large_mu_nlpc_pipeline` is a English model originally trained by MU-NLPC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/calc_baseline_t5_large_mu_nlpc_pipeline_en_5.4.2_3.0_1723953678400.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/calc_baseline_t5_large_mu_nlpc_pipeline_en_5.4.2_3.0_1723953678400.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("calc_baseline_t5_large_mu_nlpc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("calc_baseline_t5_large_mu_nlpc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|calc_baseline_t5_large_mu_nlpc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/MU-NLPC/calc-baseline-t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cendol_mt5_base_inst_id.md b/docs/_posts/ahmedlone127/2024-08-18-cendol_mt5_base_inst_id.md new file mode 100644 index 00000000000000..642b7e99956e0e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cendol_mt5_base_inst_id.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Indonesian cendol_mt5_base_inst T5Transformer from indonlp +author: John Snow Labs +name: cendol_mt5_base_inst +date: 2024-08-18 +tags: [id, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: id +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cendol_mt5_base_inst` is a Indonesian model originally trained by indonlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cendol_mt5_base_inst_id_5.4.2_3.0_1724020550243.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cendol_mt5_base_inst_id_5.4.2_3.0_1724020550243.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cendol_mt5_base_inst","id") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cendol_mt5_base_inst", "id") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cendol_mt5_base_inst| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|id| +|Size:|2.5 GB| + +## References + +https://huggingface.co/indonlp/cendol-mt5-base-inst \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_en.md b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_en.md new file mode 100644 index 00000000000000..5a6dd1502af530 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English checkpoint_22500_epoch_20 T5Transformer from pfactorial +author: John Snow Labs +name: checkpoint_22500_epoch_20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_22500_epoch_20` is a English model originally trained by pfactorial. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_22500_epoch_20_en_5.4.2_3.0_1723961664538.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_22500_epoch_20_en_5.4.2_3.0_1723961664538.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("checkpoint_22500_epoch_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("checkpoint_22500_epoch_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_22500_epoch_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pfactorial/checkpoint-22500-epoch-20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_pipeline_en.md new file mode 100644 index 00000000000000..e091c261fe9eaf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_22500_epoch_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English checkpoint_22500_epoch_20_pipeline pipeline T5Transformer from pfactorial +author: John Snow Labs +name: checkpoint_22500_epoch_20_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_22500_epoch_20_pipeline` is a English model originally trained by pfactorial. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_22500_epoch_20_pipeline_en_5.4.2_3.0_1723961713476.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_22500_epoch_20_pipeline_en_5.4.2_3.0_1723961713476.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("checkpoint_22500_epoch_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("checkpoint_22500_epoch_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_22500_epoch_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pfactorial/checkpoint-22500-epoch-20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_en.md b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_en.md new file mode 100644 index 00000000000000..e8dfd410be5990 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English checkpoint_6500_finetuned2_2010_2016 T5Transformer from kowsiknd +author: John Snow Labs +name: checkpoint_6500_finetuned2_2010_2016 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_6500_finetuned2_2010_2016` is a English model originally trained by kowsiknd. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_6500_finetuned2_2010_2016_en_5.4.2_3.0_1723962523360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_6500_finetuned2_2010_2016_en_5.4.2_3.0_1723962523360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("checkpoint_6500_finetuned2_2010_2016","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("checkpoint_6500_finetuned2_2010_2016", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_6500_finetuned2_2010_2016| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/kowsiknd/checkpoint-6500-finetuned2-2010-2016 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_pipeline_en.md new file mode 100644 index 00000000000000..1a9aeb6d6e8315 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-checkpoint_6500_finetuned2_2010_2016_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English checkpoint_6500_finetuned2_2010_2016_pipeline pipeline T5Transformer from kowsiknd +author: John Snow Labs +name: checkpoint_6500_finetuned2_2010_2016_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`checkpoint_6500_finetuned2_2010_2016_pipeline` is a English model originally trained by kowsiknd. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/checkpoint_6500_finetuned2_2010_2016_pipeline_en_5.4.2_3.0_1723962671082.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/checkpoint_6500_finetuned2_2010_2016_pipeline_en_5.4.2_3.0_1723962671082.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("checkpoint_6500_finetuned2_2010_2016_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("checkpoint_6500_finetuned2_2010_2016_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|checkpoint_6500_finetuned2_2010_2016_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/kowsiknd/checkpoint-6500-finetuned2-2010-2016 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_es.md b/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_es.md new file mode 100644 index 00000000000000..0529264a25c9a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish cleaner_restaurant_names T5Transformer from felipeace96 +author: John Snow Labs +name: cleaner_restaurant_names +date: 2024-08-18 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cleaner_restaurant_names` is a Castilian, Spanish model originally trained by felipeace96. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cleaner_restaurant_names_es_5.4.2_3.0_1723949763944.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cleaner_restaurant_names_es_5.4.2_3.0_1723949763944.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cleaner_restaurant_names","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cleaner_restaurant_names", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cleaner_restaurant_names| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|932.3 MB| + +## References + +https://huggingface.co/felipeace96/cleaner-restaurant-names \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_pipeline_es.md new file mode 100644 index 00000000000000..c45ad29e8447f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cleaner_restaurant_names_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish cleaner_restaurant_names_pipeline pipeline T5Transformer from felipeace96 +author: John Snow Labs +name: cleaner_restaurant_names_pipeline +date: 2024-08-18 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cleaner_restaurant_names_pipeline` is a Castilian, Spanish model originally trained by felipeace96. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cleaner_restaurant_names_pipeline_es_5.4.2_3.0_1723949838176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cleaner_restaurant_names_pipeline_es_5.4.2_3.0_1723949838176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cleaner_restaurant_names_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cleaner_restaurant_names_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cleaner_restaurant_names_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|932.3 MB| + +## References + +https://huggingface.co/felipeace96/cleaner-restaurant-names + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_en.md b/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_en.md new file mode 100644 index 00000000000000..e8c2c6a3ea85f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English codetranslargetfflt T5Transformer from AlexC98 +author: John Snow Labs +name: codetranslargetfflt +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`codetranslargetfflt` is a English model originally trained by AlexC98. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/codetranslargetfflt_en_5.4.2_3.0_1723983111383.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/codetranslargetfflt_en_5.4.2_3.0_1723983111383.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("codetranslargetfflt","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("codetranslargetfflt", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|codetranslargetfflt| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/AlexC98/CodeTransLargeTFFlt \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_pipeline_en.md new file mode 100644 index 00000000000000..80bf09f3d1aa11 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-codetranslargetfflt_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English codetranslargetfflt_pipeline pipeline T5Transformer from AlexC98 +author: John Snow Labs +name: codetranslargetfflt_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`codetranslargetfflt_pipeline` is a English model originally trained by AlexC98. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/codetranslargetfflt_pipeline_en_5.4.2_3.0_1723983256031.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/codetranslargetfflt_pipeline_en_5.4.2_3.0_1723983256031.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("codetranslargetfflt_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("codetranslargetfflt_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|codetranslargetfflt_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/AlexC98/CodeTransLargeTFFlt + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_pipeline_zh.md b/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_pipeline_zh.md new file mode 100644 index 00000000000000..68937d728998ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_pipeline_zh.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Chinese comet_atomic_chinese_pipeline pipeline T5Transformer from svjack +author: John Snow Labs +name: comet_atomic_chinese_pipeline +date: 2024-08-18 +tags: [zh, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`comet_atomic_chinese_pipeline` is a Chinese model originally trained by svjack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/comet_atomic_chinese_pipeline_zh_5.4.2_3.0_1724013316600.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/comet_atomic_chinese_pipeline_zh_5.4.2_3.0_1724013316600.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("comet_atomic_chinese_pipeline", lang = "zh") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("comet_atomic_chinese_pipeline", lang = "zh") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|comet_atomic_chinese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|zh| +|Size:|954.3 MB| + +## References + +https://huggingface.co/svjack/comet-atomic-zh + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_zh.md b/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_zh.md new file mode 100644 index 00000000000000..ab78de856851eb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-comet_atomic_chinese_zh.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Chinese comet_atomic_chinese T5Transformer from svjack +author: John Snow Labs +name: comet_atomic_chinese +date: 2024-08-18 +tags: [zh, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`comet_atomic_chinese` is a Chinese model originally trained by svjack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/comet_atomic_chinese_zh_5.4.2_3.0_1724013265411.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/comet_atomic_chinese_zh_5.4.2_3.0_1724013265411.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("comet_atomic_chinese","zh") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("comet_atomic_chinese", "zh") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|comet_atomic_chinese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|zh| +|Size:|954.3 MB| + +## References + +https://huggingface.co/svjack/comet-atomic-zh \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_en.md new file mode 100644 index 00000000000000..83be181d32ce2a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting10_aspol_vcheck T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting10_aspol_vcheck +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting10_aspol_vcheck` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting10_aspol_vcheck_en_5.4.2_3.0_1723939327788.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting10_aspol_vcheck_en_5.4.2_3.0_1723939327788.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting10_aspol_vcheck","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting10_aspol_vcheck", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting10_aspol_vcheck| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting10_ASPOL_vcheck \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline_en.md new file mode 100644 index 00000000000000..b675d98eaf369c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline_en_5.4.2_3.0_1723939518719.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline_en_5.4.2_3.0_1723939518719.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting10_aspol_vcheck_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting10_ASPOL_vcheck + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_eql_aug3_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_eql_aug3_en.md new file mode 100644 index 00000000000000..6c9ce14df05501 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_eql_aug3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_eql_aug3 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_eql_aug3 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_eql_aug3` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_eql_aug3_en_5.4.2_3.0_1723956693807.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_eql_aug3_en_5.4.2_3.0_1723956693807.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_eql_aug3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_eql_aug3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_eql_aug3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_EQL_Aug3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_en.md new file mode 100644 index 00000000000000..955e13e99169dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_en_5.4.2_3.0_1723996153674.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_en_5.4.2_3.0_1723996153674.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_GPT_T3vsLabel2T5_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline_en.md new file mode 100644 index 00000000000000..2fac2dbd06f118 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline_en_5.4.2_3.0_1723996323014.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline_en_5.4.2_3.0_1723996323014.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_gpt_t3vslabel2t5_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_GPT_T3vsLabel2T5_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_en.md new file mode 100644 index 00000000000000..98bf8f239a367a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_label2text_ap1 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_label2text_ap1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_label2text_ap1` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_label2text_ap1_en_5.4.2_3.0_1724005387048.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_label2text_ap1_en_5.4.2_3.0_1724005387048.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_label2text_ap1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_label2text_ap1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_label2text_ap1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_label2text_ap1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline_en.md new file mode 100644 index 00000000000000..0594be5569eb44 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline_en_5.4.2_3.0_1724005554039.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline_en_5.4.2_3.0_1724005554039.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_label2text_ap1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_label2text_ap1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_en.md new file mode 100644 index 00000000000000..12a2cb167dab82 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_sup_augfull T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_sup_augfull +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_sup_augfull` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_augfull_en_5.4.2_3.0_1723992299021.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_augfull_en_5.4.2_3.0_1723992299021.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_sup_augfull","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_apsol_sup_augfull", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_sup_augfull| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_SUP_AugFull \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline_en.md new file mode 100644 index 00000000000000..6cdd67306f7686 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline_en_5.4.2_3.0_1723992478122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline_en_5.4.2_3.0_1723992478122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_apsol_sup_augfull_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_APSOL_SUP_AugFull + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_en.md new file mode 100644 index 00000000000000..d018bf28421083 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_aspol_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_aspol_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_aspol_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_v2_en_5.4.2_3.0_1723980048329.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_v2_en_5.4.2_3.0_1723980048329.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_aspol_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_aspol_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_aspol_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_ASPOL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_pipeline_en.md new file mode 100644 index 00000000000000..b5f67171a518d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_aspol_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_aspol_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_aspol_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_aspol_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_v2_pipeline_en_5.4.2_3.0_1723980215333.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_v2_pipeline_en_5.4.2_3.0_1723980215333.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_aspol_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_aspol_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_aspol_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_ASPOL_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_en.md new file mode 100644 index 00000000000000..c608293e81fa52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_spaol T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_spaol +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_spaol` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_spaol_en_5.4.2_3.0_1723993551156.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_spaol_en_5.4.2_3.0_1723993551156.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_spaol","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_spaol", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_spaol| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_SPAOL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_pipeline_en.md new file mode 100644 index 00000000000000..ff6a02fd778f0f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_prompting5_spaol_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_spaol_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_spaol_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_spaol_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_spaol_pipeline_en_5.4.2_3.0_1723993716649.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_spaol_pipeline_en_5.4.2_3.0_1723993716649.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_spaol_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_spaol_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_spaol_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_SPAOL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_en.md new file mode 100644 index 00000000000000..ee4a04771fb9cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_apsol_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_apsol_v1_h1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_apsol_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_apsol_v1_h1_en_5.4.2_3.0_1723951973295.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_apsol_v1_h1_en_5.4.2_3.0_1723951973295.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_apsol_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_apsol_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_apsol_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_APSOL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..a7b0bd079a69b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline_en_5.4.2_3.0_1723952146958.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline_en_5.4.2_3.0_1723952146958.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_apsol_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_APSOL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_en.md new file mode 100644 index 00000000000000..766859d1e61221 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_poasl_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_poasl_v1_h1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_poasl_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_poasl_v1_h1_en_5.4.2_3.0_1723942396182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_poasl_v1_h1_en_5.4.2_3.0_1723942396182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_poasl_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_poasl_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_poasl_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_POASL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..2b267ead60e999 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline_en_5.4.2_3.0_1723942583763.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline_en_5.4.2_3.0_1723942583763.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_poasl_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_POASL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction4_aopsl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction4_aopsl_v1_en.md new file mode 100644 index 00000000000000..59f926f48cd628 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instruction4_aopsl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction4_aopsl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction4_aopsl_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction4_aopsl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction4_aopsl_v1_en_5.4.2_3.0_1723977296232.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction4_aopsl_v1_en_5.4.2_3.0_1723977296232.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction4_aopsl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction4_aopsl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction4_aopsl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction4_AOPSL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_en.md new file mode 100644 index 00000000000000..31e2474c796ec9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn0_soapl_v1_h0 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn0_soapl_v1_h0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn0_soapl_v1_h0` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_en_5.4.2_3.0_1723974056033.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_en_5.4.2_3.0_1723974056033.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn0_soapl_v1_h0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instructionn0_soapl_v1_h0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn0_soapl_v1_h0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN0_SOAPL_v1_h0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline_en.md new file mode 100644 index 00000000000000..a65839a8571e7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline_en_5.4.2_3.0_1723974220543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline_en_5.4.2_3.0_1723974220543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instructionn0_soapl_v1_h0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_InstructionN0_SOAPL_v1_h0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_en.md new file mode 100644 index 00000000000000..0f184d8606c418 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_poasl T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_poasl +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_poasl` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_poasl_en_5.4.2_3.0_1724008024786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_poasl_en_5.4.2_3.0_1724008024786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_poasl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_poasl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_poasl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_POASL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_pipeline_en.md new file mode 100644 index 00000000000000..df7a54d42c7870 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_poasl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_poasl_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_poasl_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_poasl_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_poasl_pipeline_en_5.4.2_3.0_1724008188620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_poasl_pipeline_en_5.4.2_3.0_1724008188620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_poasl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_poasl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_poasl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_POASL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_posal_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_posal_v1_en.md new file mode 100644 index 00000000000000..419dc9eadbd4bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_posal_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_posal_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_posal_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_posal_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_posal_v1_en_5.4.2_3.0_1723960604442.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_posal_v1_en_5.4.2_3.0_1723960604442.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_posal_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_posal_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_posal_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_POSAL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_en.md new file mode 100644 index 00000000000000..52d5f932ba4a92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_psaol T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_psaol +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_psaol` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_en_5.4.2_3.0_1723978310596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_en_5.4.2_3.0_1723978310596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_psaol","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_psaol", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_psaol| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_PSAOL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_en.md new file mode 100644 index 00000000000000..89f5cbadb5554a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_psaol_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_psaol_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_psaol_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_v1_en_5.4.2_3.0_1723944086741.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_v1_en_5.4.2_3.0_1723944086741.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_psaol_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_psaol_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_psaol_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_PSAOL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline_en.md new file mode 100644 index 00000000000000..c1f65df975575a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline_en_5.4.2_3.0_1723944262537.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline_en_5.4.2_3.0_1723944262537.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_psaol_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_PSAOL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_sopal_h1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_sopal_h1_en.md new file mode 100644 index 00000000000000..8bbd3b96333386 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_sopal_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_sopal_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_sopal_h1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_sopal_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sopal_h1_en_5.4.2_3.0_1723961123860.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sopal_h1_en_5.4.2_3.0_1723961123860.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_sopal_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_sopal_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_sopal_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SOPAL_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_en.md new file mode 100644 index 00000000000000..9ad40f4d741d7d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_spoal_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_spoal_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_spoal_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spoal_v1_en_5.4.2_3.0_1723955153062.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spoal_v1_en_5.4.2_3.0_1723955153062.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_spoal_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_spoal_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_spoal_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SPOAL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline_en.md new file mode 100644 index 00000000000000..744ce204d18b5a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline_en_5.4.2_3.0_1723955332533.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline_en_5.4.2_3.0_1723955332533.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_spoal_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SPOAL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_en.md new file mode 100644 index 00000000000000..6eae237f7bd422 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction3_soapl T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction3_soapl +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction3_soapl` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_en_5.4.2_3.0_1723996525507.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_en_5.4.2_3.0_1723996525507.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction3_soapl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction3_soapl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction3_soapl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction3_SOAPL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_pipeline_en.md new file mode 100644 index 00000000000000..1d833ebd23d4f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instruction3_soapl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction3_soapl_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction3_soapl_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction3_soapl_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_pipeline_en_5.4.2_3.0_1723996685979.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_pipeline_en_5.4.2_3.0_1723996685979.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction3_soapl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction3_soapl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction3_soapl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction3_SOAPL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_en.md new file mode 100644 index 00000000000000..d2defb861299fd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn0_saopl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn0_saopl_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn0_saopl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_saopl_v1_en_5.4.2_3.0_1723966496095.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_saopl_v1_en_5.4.2_3.0_1723966496095.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn0_saopl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn0_saopl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn0_saopl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN0_SAOPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline_en.md new file mode 100644 index 00000000000000..fc697878b4eb66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline_en_5.4.2_3.0_1723966675684.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline_en_5.4.2_3.0_1723966675684.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn0_saopl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN0_SAOPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_en.md new file mode 100644 index 00000000000000..54b163cba1133e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_paosl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_paosl_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_paosl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_paosl_v1_en_5.4.2_3.0_1723945667389.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_paosl_v1_en_5.4.2_3.0_1723945667389.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_paosl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_paosl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_paosl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_PAOSL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline_en.md new file mode 100644 index 00000000000000..e66dca6d22484f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline_en_5.4.2_3.0_1723945856019.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline_en_5.4.2_3.0_1723945856019.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_paosl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_PAOSL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_en.md new file mode 100644 index 00000000000000..65d75b80e8722a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_poasl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_poasl_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_poasl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_poasl_v1_en_5.4.2_3.0_1723947600880.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_poasl_v1_en_5.4.2_3.0_1723947600880.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_poasl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_poasl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_poasl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_POASL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline_en.md new file mode 100644 index 00000000000000..d514305dbf06fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline_en_5.4.2_3.0_1723947777686.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline_en_5.4.2_3.0_1723947777686.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_poasl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_POASL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_en.md b/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_en.md new file mode 100644 index 00000000000000..b065f8ab832384 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English custom_billsum_model_dead_owwl T5Transformer from dead-owwl +author: John Snow Labs +name: custom_billsum_model_dead_owwl +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`custom_billsum_model_dead_owwl` is a English model originally trained by dead-owwl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/custom_billsum_model_dead_owwl_en_5.4.2_3.0_1723996937876.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/custom_billsum_model_dead_owwl_en_5.4.2_3.0_1723996937876.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("custom_billsum_model_dead_owwl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("custom_billsum_model_dead_owwl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|custom_billsum_model_dead_owwl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|323.1 MB| + +## References + +https://huggingface.co/dead-owwl/custom_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_pipeline_en.md new file mode 100644 index 00000000000000..ae967876934dd9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-custom_billsum_model_dead_owwl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English custom_billsum_model_dead_owwl_pipeline pipeline T5Transformer from dead-owwl +author: John Snow Labs +name: custom_billsum_model_dead_owwl_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`custom_billsum_model_dead_owwl_pipeline` is a English model originally trained by dead-owwl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/custom_billsum_model_dead_owwl_pipeline_en_5.4.2_3.0_1723996955660.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/custom_billsum_model_dead_owwl_pipeline_en_5.4.2_3.0_1723996955660.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("custom_billsum_model_dead_owwl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("custom_billsum_model_dead_owwl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|custom_billsum_model_dead_owwl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|323.1 MB| + +## References + +https://huggingface.co/dead-owwl/custom_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_en.md b/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_en.md new file mode 100644 index 00000000000000..66762693dcb1df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English descriptions_500k_top10k_llm T5Transformer from bitadin +author: John Snow Labs +name: descriptions_500k_top10k_llm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`descriptions_500k_top10k_llm` is a English model originally trained by bitadin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/descriptions_500k_top10k_llm_en_5.4.2_3.0_1723944708914.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/descriptions_500k_top10k_llm_en_5.4.2_3.0_1723944708914.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("descriptions_500k_top10k_llm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("descriptions_500k_top10k_llm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|descriptions_500k_top10k_llm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bitadin/descriptions-500k-top10k-llm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_pipeline_en.md new file mode 100644 index 00000000000000..7b3bfbc22810b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-descriptions_500k_top10k_llm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English descriptions_500k_top10k_llm_pipeline pipeline T5Transformer from bitadin +author: John Snow Labs +name: descriptions_500k_top10k_llm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`descriptions_500k_top10k_llm_pipeline` is a English model originally trained by bitadin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/descriptions_500k_top10k_llm_pipeline_en_5.4.2_3.0_1723944759904.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/descriptions_500k_top10k_llm_pipeline_en_5.4.2_3.0_1723944759904.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("descriptions_500k_top10k_llm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("descriptions_500k_top10k_llm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|descriptions_500k_top10k_llm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bitadin/descriptions-500k-top10k-llm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_en.md new file mode 100644 index 00000000000000..c7263c52a65273 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_010099_0_5 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_010099_0_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_010099_0_5` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_0_5_en_5.4.2_3.0_1723964813564.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_0_5_en_5.4.2_3.0_1723964813564.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_010099_0_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_010099_0_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_010099_0_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-010099-0.5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_pipeline_en.md new file mode 100644 index 00000000000000..14b67986ff4433 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_010099_0_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_010099_0_5_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_010099_0_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_010099_0_5_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_0_5_pipeline_en_5.4.2_3.0_1723965007053.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_0_5_pipeline_en_5.4.2_3.0_1723965007053.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_010099_0_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_010099_0_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_010099_0_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-010099-0.5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_en.md new file mode 100644 index 00000000000000..cc484502af0c09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_0_03_1 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_03_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_03_1` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_03_1_en_5.4.2_3.0_1723984934998.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_03_1_en_5.4.2_3.0_1723984934998.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_0_03_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_0_03_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_03_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.03-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_pipeline_en.md new file mode 100644 index 00000000000000..cfc1da90c9c4da --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-distilled_mt5_small_0_03_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_0_03_1_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_03_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_03_1_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_03_1_pipeline_en_5.4.2_3.0_1723985120268.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_03_1_pipeline_en_5.4.2_3.0_1723985120268.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_0_03_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_0_03_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_03_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.03-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_bn.md b/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_bn.md new file mode 100644 index 00000000000000..9e33d1d087937d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_bn.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Bengali doc2query_banglat5_indicmarco T5Transformer from iutsslir +author: John Snow Labs +name: doc2query_banglat5_indicmarco +date: 2024-08-18 +tags: [bn, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`doc2query_banglat5_indicmarco` is a Bengali model originally trained by iutsslir. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/doc2query_banglat5_indicmarco_bn_5.4.2_3.0_1723983036820.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/doc2query_banglat5_indicmarco_bn_5.4.2_3.0_1723983036820.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("doc2query_banglat5_indicmarco","bn") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("doc2query_banglat5_indicmarco", "bn") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|doc2query_banglat5_indicmarco| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|bn| +|Size:|1.0 GB| + +## References + +https://huggingface.co/iutsslir/doc2query-banglat5-indicmarco \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_pipeline_bn.md b/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_pipeline_bn.md new file mode 100644 index 00000000000000..c1cd6e6ed304cc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-doc2query_banglat5_indicmarco_pipeline_bn.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Bengali doc2query_banglat5_indicmarco_pipeline pipeline T5Transformer from iutsslir +author: John Snow Labs +name: doc2query_banglat5_indicmarco_pipeline +date: 2024-08-18 +tags: [bn, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: bn +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`doc2query_banglat5_indicmarco_pipeline` is a Bengali model originally trained by iutsslir. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/doc2query_banglat5_indicmarco_pipeline_bn_5.4.2_3.0_1723983088076.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/doc2query_banglat5_indicmarco_pipeline_bn_5.4.2_3.0_1723983088076.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("doc2query_banglat5_indicmarco_pipeline", lang = "bn") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("doc2query_banglat5_indicmarco_pipeline", lang = "bn") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|doc2query_banglat5_indicmarco_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|bn| +|Size:|1.0 GB| + +## References + +https://huggingface.co/iutsslir/doc2query-banglat5-indicmarco + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_en.md b/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_en.md new file mode 100644 index 00000000000000..59519fecb4f9f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English doc2query_ppo_msmarco_8192_121 T5Transformer from Hermi2023 +author: John Snow Labs +name: doc2query_ppo_msmarco_8192_121 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`doc2query_ppo_msmarco_8192_121` is a English model originally trained by Hermi2023. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/doc2query_ppo_msmarco_8192_121_en_5.4.2_3.0_1723939236454.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/doc2query_ppo_msmarco_8192_121_en_5.4.2_3.0_1723939236454.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("doc2query_ppo_msmarco_8192_121","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("doc2query_ppo_msmarco_8192_121", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|doc2query_ppo_msmarco_8192_121| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/Hermi2023/doc2query-ppo-msmarco-8192-121 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_pipeline_en.md new file mode 100644 index 00000000000000..e6ffca778b3b45 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-doc2query_ppo_msmarco_8192_121_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English doc2query_ppo_msmarco_8192_121_pipeline pipeline T5Transformer from Hermi2023 +author: John Snow Labs +name: doc2query_ppo_msmarco_8192_121_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`doc2query_ppo_msmarco_8192_121_pipeline` is a English model originally trained by Hermi2023. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/doc2query_ppo_msmarco_8192_121_pipeline_en_5.4.2_3.0_1723939289853.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/doc2query_ppo_msmarco_8192_121_pipeline_en_5.4.2_3.0_1723939289853.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("doc2query_ppo_msmarco_8192_121_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("doc2query_ppo_msmarco_8192_121_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|doc2query_ppo_msmarco_8192_121_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/Hermi2023/doc2query-ppo-msmarco-8192-121 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_en.md new file mode 100644 index 00000000000000..c638343e697664 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English email_extractor_more_data_model_t5_small T5Transformer from cmtn +author: John Snow Labs +name: email_extractor_more_data_model_t5_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`email_extractor_more_data_model_t5_small` is a English model originally trained by cmtn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/email_extractor_more_data_model_t5_small_en_5.4.2_3.0_1724002975571.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/email_extractor_more_data_model_t5_small_en_5.4.2_3.0_1724002975571.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("email_extractor_more_data_model_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("email_extractor_more_data_model_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|email_extractor_more_data_model_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|300.8 MB| + +## References + +https://huggingface.co/cmtn/email_extractor_more_data_model_t5_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..abff9c6f84aeff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-email_extractor_more_data_model_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English email_extractor_more_data_model_t5_small_pipeline pipeline T5Transformer from cmtn +author: John Snow Labs +name: email_extractor_more_data_model_t5_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`email_extractor_more_data_model_t5_small_pipeline` is a English model originally trained by cmtn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/email_extractor_more_data_model_t5_small_pipeline_en_5.4.2_3.0_1724003000380.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/email_extractor_more_data_model_t5_small_pipeline_en_5.4.2_3.0_1724003000380.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("email_extractor_more_data_model_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("email_extractor_more_data_model_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|email_extractor_more_data_model_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|300.8 MB| + +## References + +https://huggingface.co/cmtn/email_extractor_more_data_model_t5_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_en.md new file mode 100644 index 00000000000000..3b69df9474b8a2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_keyword_all_ptr_normal T5Transformer from NanaEilish +author: John Snow Labs +name: english_keyword_all_ptr_normal +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_keyword_all_ptr_normal` is a English model originally trained by NanaEilish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_keyword_all_ptr_normal_en_5.4.2_3.0_1724011226436.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_keyword_all_ptr_normal_en_5.4.2_3.0_1724011226436.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_keyword_all_ptr_normal","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_keyword_all_ptr_normal", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_keyword_all_ptr_normal| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|976.5 MB| + +## References + +https://huggingface.co/NanaEilish/en_keyword_all_ptr_normal \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_pipeline_en.md new file mode 100644 index 00000000000000..b7cd93229df157 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_keyword_all_ptr_normal_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_keyword_all_ptr_normal_pipeline pipeline T5Transformer from NanaEilish +author: John Snow Labs +name: english_keyword_all_ptr_normal_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_keyword_all_ptr_normal_pipeline` is a English model originally trained by NanaEilish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_keyword_all_ptr_normal_pipeline_en_5.4.2_3.0_1724011273934.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_keyword_all_ptr_normal_pipeline_en_5.4.2_3.0_1724011273934.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_keyword_all_ptr_normal_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_keyword_all_ptr_normal_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_keyword_all_ptr_normal_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|976.5 MB| + +## References + +https://huggingface.co/NanaEilish/en_keyword_all_ptr_normal + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_en.md new file mode 100644 index 00000000000000..d99a349853ed09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_mt5_small_5_spider_e22vvb T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_small_5_spider_e22vvb +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_small_5_spider_e22vvb` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_small_5_spider_e22vvb_en_5.4.2_3.0_1723983565299.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_small_5_spider_e22vvb_en_5.4.2_3.0_1723983565299.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_mt5_small_5_spider_e22vvb","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_mt5_small_5_spider_e22vvb", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_small_5_spider_e22vvb| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-small_5_spider \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_pipeline_en.md new file mode 100644 index 00000000000000..49eb90329fe026 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_mt5_small_5_spider_e22vvb_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_mt5_small_5_spider_e22vvb_pipeline pipeline T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_small_5_spider_e22vvb_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_small_5_spider_e22vvb_pipeline` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_small_5_spider_e22vvb_pipeline_en_5.4.2_3.0_1723983762305.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_small_5_spider_e22vvb_pipeline_en_5.4.2_3.0_1723983762305.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_mt5_small_5_spider_e22vvb_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_mt5_small_5_spider_e22vvb_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_small_5_spider_e22vvb_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-small_5_spider + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_en.md new file mode 100644 index 00000000000000..2a611a438bd3d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_vietnamese_envit5_translation_doc_train T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_envit5_translation_doc_train +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_envit5_translation_doc_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_translation_doc_train_en_5.4.2_3.0_1723988538614.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_translation_doc_train_en_5.4.2_3.0_1723988538614.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_vietnamese_envit5_translation_doc_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_vietnamese_envit5_translation_doc_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_envit5_translation_doc_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_envit5-translation_doc_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_pipeline_en.md new file mode 100644 index 00000000000000..66531415661d75 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-english_vietnamese_envit5_translation_doc_train_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_vietnamese_envit5_translation_doc_train_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_envit5_translation_doc_train_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_envit5_translation_doc_train_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_translation_doc_train_pipeline_en_5.4.2_3.0_1723988613838.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_translation_doc_train_pipeline_en_5.4.2_3.0_1723988613838.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_vietnamese_envit5_translation_doc_train_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_vietnamese_envit5_translation_doc_train_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_envit5_translation_doc_train_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_envit5-translation_doc_train + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_en.md b/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_en.md new file mode 100644 index 00000000000000..2e0f919a663e59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ensemble_icl_t5_lm_large T5Transformer from qinyuany +author: John Snow Labs +name: ensemble_icl_t5_lm_large +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ensemble_icl_t5_lm_large` is a English model originally trained by qinyuany. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ensemble_icl_t5_lm_large_en_5.4.2_3.0_1723990079134.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ensemble_icl_t5_lm_large_en_5.4.2_3.0_1723990079134.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ensemble_icl_t5_lm_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ensemble_icl_t5_lm_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ensemble_icl_t5_lm_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/qinyuany/ensemble-icl-t5-lm-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_pipeline_en.md new file mode 100644 index 00000000000000..a7748d6b132a84 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ensemble_icl_t5_lm_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ensemble_icl_t5_lm_large_pipeline pipeline T5Transformer from qinyuany +author: John Snow Labs +name: ensemble_icl_t5_lm_large_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ensemble_icl_t5_lm_large_pipeline` is a English model originally trained by qinyuany. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ensemble_icl_t5_lm_large_pipeline_en_5.4.2_3.0_1723990213749.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ensemble_icl_t5_lm_large_pipeline_en_5.4.2_3.0_1723990213749.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ensemble_icl_t5_lm_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ensemble_icl_t5_lm_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ensemble_icl_t5_lm_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/qinyuany/ensemble-icl-t5-lm-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-entailer_large_en.md b/docs/_posts/ahmedlone127/2024-08-18-entailer_large_en.md new file mode 100644 index 00000000000000..d64a0ecd81d38a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-entailer_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English entailer_large T5Transformer from allenai +author: John Snow Labs +name: entailer_large +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`entailer_large` is a English model originally trained by allenai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/entailer_large_en_5.4.2_3.0_1724017942340.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/entailer_large_en_5.4.2_3.0_1724017942340.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("entailer_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("entailer_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|entailer_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/allenai/entailer-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-entailer_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-entailer_large_pipeline_en.md new file mode 100644 index 00000000000000..d438b24033b7a8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-entailer_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English entailer_large_pipeline pipeline T5Transformer from allenai +author: John Snow Labs +name: entailer_large_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`entailer_large_pipeline` is a English model originally trained by allenai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/entailer_large_pipeline_en_5.4.2_3.0_1724018426118.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/entailer_large_pipeline_en_5.4.2_3.0_1724018426118.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("entailer_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("entailer_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|entailer_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/allenai/entailer-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_en.md new file mode 100644 index 00000000000000..d921b1928e16bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1 T5Transformer from harish +author: John Snow Labs +name: esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_en_5.4.2_3.0_1724002752587.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_en_5.4.2_3.0_1724002752587.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/harish/eSNLI-limited-eFigSNLI-e10-a0-9-eFigSNLI-e20-a-0-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline_en.md new file mode 100644 index 00000000000000..01fde7136cccb8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline pipeline T5Transformer from harish +author: John Snow Labs +name: esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline_en_5.4.2_3.0_1724002801847.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline_en_5.4.2_3.0_1724002801847.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|esnli_limited_efigsnli_e10_a0_9_efigsnli_e20_a_0_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/harish/eSNLI-limited-eFigSNLI-e10-a0-9-eFigSNLI-e20-a-0-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_en.md b/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_en.md new file mode 100644 index 00000000000000..668f1a560bd5d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English expected_model_nov11_tanvirsrbd1 T5Transformer from tanvirsrbd1 +author: John Snow Labs +name: expected_model_nov11_tanvirsrbd1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`expected_model_nov11_tanvirsrbd1` is a English model originally trained by tanvirsrbd1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/expected_model_nov11_tanvirsrbd1_en_5.4.2_3.0_1724014847481.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/expected_model_nov11_tanvirsrbd1_en_5.4.2_3.0_1724014847481.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("expected_model_nov11_tanvirsrbd1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("expected_model_nov11_tanvirsrbd1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|expected_model_nov11_tanvirsrbd1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tanvirsrbd1/expected_model_nov11 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_pipeline_en.md new file mode 100644 index 00000000000000..ec79488025a377 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-expected_model_nov11_tanvirsrbd1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English expected_model_nov11_tanvirsrbd1_pipeline pipeline T5Transformer from tanvirsrbd1 +author: John Snow Labs +name: expected_model_nov11_tanvirsrbd1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`expected_model_nov11_tanvirsrbd1_pipeline` is a English model originally trained by tanvirsrbd1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/expected_model_nov11_tanvirsrbd1_pipeline_en_5.4.2_3.0_1724014896956.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/expected_model_nov11_tanvirsrbd1_pipeline_en_5.4.2_3.0_1724014896956.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("expected_model_nov11_tanvirsrbd1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("expected_model_nov11_tanvirsrbd1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|expected_model_nov11_tanvirsrbd1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tanvirsrbd1/expected_model_nov11 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1_en.md new file mode 100644 index 00000000000000..3b80099514ad05 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1 T5Transformer from tau +author: John Snow Labs +name: false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1` is a English model originally trained by tau. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1_en_5.4.2_3.0_1724003064669.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1_en_5.4.2_3.0_1724003064669.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|false_large_pmi_para0_sent1_span2_itfalse_ssoftmax_rrfalse_8_1024_0_15_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/tau/False_large_pmi_para0_sent1_span2_itFalse_ssoftmax_rrFalse_8_1024_0.15_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1_en.md new file mode 100644 index 00000000000000..034a471c9f9e19 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1 T5Transformer from tau +author: John Snow Labs +name: false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1` is a English model originally trained by tau. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1_en_5.4.2_3.0_1723989269489.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1_en_5.4.2_3.0_1723989269489.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|false_large_rouge_paranone_sent0_spannone_itfalse_sargmax_rrfalse_8_1024_0_15_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/tau/False_large_rouge_paraNone_sent0_spanNone_itFalse_sargmax_rrFalse_8_1024_0.15_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-false_large_t5_lm_8_1024_0_15_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-false_large_t5_lm_8_1024_0_15_1_en.md new file mode 100644 index 00000000000000..6bb2581ee223dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-false_large_t5_lm_8_1024_0_15_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English false_large_t5_lm_8_1024_0_15_1 T5Transformer from tau +author: John Snow Labs +name: false_large_t5_lm_8_1024_0_15_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`false_large_t5_lm_8_1024_0_15_1` is a English model originally trained by tau. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/false_large_t5_lm_8_1024_0_15_1_en_5.4.2_3.0_1723963357805.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/false_large_t5_lm_8_1024_0_15_1_en_5.4.2_3.0_1723963357805.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("false_large_t5_lm_8_1024_0_15_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("false_large_t5_lm_8_1024_0_15_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|false_large_t5_lm_8_1024_0_15_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/tau/False_large_t5_lm_8_1024_0.15_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_en.md b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_en.md new file mode 100644 index 00000000000000..c4cbafdcdd4cab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English fine_tuned_bahasa_melayu_200000 T5Transformer from yuvaranianandhan24 +author: John Snow Labs +name: fine_tuned_bahasa_melayu_200000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`fine_tuned_bahasa_melayu_200000` is a English model originally trained by yuvaranianandhan24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_200000_en_5.4.2_3.0_1724000846645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_200000_en_5.4.2_3.0_1724000846645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("fine_tuned_bahasa_melayu_200000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("fine_tuned_bahasa_melayu_200000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|fine_tuned_bahasa_melayu_200000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yuvaranianandhan24/fine-tuned-bahasa-melayu-200000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_pipeline_en.md new file mode 100644 index 00000000000000..5ef2fdd1528b07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_200000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English fine_tuned_bahasa_melayu_200000_pipeline pipeline T5Transformer from yuvaranianandhan24 +author: John Snow Labs +name: fine_tuned_bahasa_melayu_200000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`fine_tuned_bahasa_melayu_200000_pipeline` is a English model originally trained by yuvaranianandhan24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_200000_pipeline_en_5.4.2_3.0_1724000893112.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_200000_pipeline_en_5.4.2_3.0_1724000893112.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("fine_tuned_bahasa_melayu_200000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("fine_tuned_bahasa_melayu_200000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|fine_tuned_bahasa_melayu_200000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yuvaranianandhan24/fine-tuned-bahasa-melayu-200000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_en.md b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_en.md new file mode 100644 index 00000000000000..1b6d41dd6518f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English fine_tuned_bahasa_melayu_2l T5Transformer from PiraiAI +author: John Snow Labs +name: fine_tuned_bahasa_melayu_2l +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`fine_tuned_bahasa_melayu_2l` is a English model originally trained by PiraiAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_2l_en_5.4.2_3.0_1723942946625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_2l_en_5.4.2_3.0_1723942946625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("fine_tuned_bahasa_melayu_2l","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("fine_tuned_bahasa_melayu_2l", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|fine_tuned_bahasa_melayu_2l| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PiraiAI/fine-tuned-bahasa-melayu-2L \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_pipeline_en.md new file mode 100644 index 00000000000000..b9a990aa130ca2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-fine_tuned_bahasa_melayu_2l_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English fine_tuned_bahasa_melayu_2l_pipeline pipeline T5Transformer from PiraiAI +author: John Snow Labs +name: fine_tuned_bahasa_melayu_2l_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`fine_tuned_bahasa_melayu_2l_pipeline` is a English model originally trained by PiraiAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_2l_pipeline_en_5.4.2_3.0_1723942996597.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/fine_tuned_bahasa_melayu_2l_pipeline_en_5.4.2_3.0_1723942996597.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("fine_tuned_bahasa_melayu_2l_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("fine_tuned_bahasa_melayu_2l_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|fine_tuned_bahasa_melayu_2l_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PiraiAI/fine-tuned-bahasa-melayu-2L + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_en.md new file mode 100644 index 00000000000000..7e736e933f8e6f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_1_6_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_6_cnndm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_6_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_cnndm_en_5.4.2_3.0_1723994625053.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_cnndm_en_5.4.2_3.0_1723994625053.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_1_6_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_1_6_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_6_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|754.0 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-6-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..6f321e7b92f86f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_1_6_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_1_6_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_6_cnndm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_6_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_cnndm_pipeline_en_5.4.2_3.0_1723994659253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_cnndm_pipeline_en_5.4.2_3.0_1723994659253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_1_6_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_1_6_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_6_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|754.0 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-6-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_en.md new file mode 100644 index 00000000000000..2c8d920d3a30e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_2_2_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_2_2_cnndm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_2_2_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_2_2_cnndm_en_5.4.2_3.0_1723983058014.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_2_2_cnndm_en_5.4.2_3.0_1723983058014.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_2_2_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_2_2_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_2_2_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|524.2 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-2-2-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..e75e2b048b2d21 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_2_2_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_2_2_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_2_2_cnndm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_2_2_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_2_2_cnndm_pipeline_en_5.4.2_3.0_1723983082760.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_2_2_cnndm_pipeline_en_5.4.2_3.0_1723983082760.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_2_2_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_2_2_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_2_2_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|524.2 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-2-2-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_en.md new file mode 100644 index 00000000000000..414e846886d98b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_6_4_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_6_4_cnndm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_6_4_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_6_4_cnndm_en_5.4.2_3.0_1724004808594.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_6_4_cnndm_en_5.4.2_3.0_1724004808594.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_6_4_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_6_4_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_6_4_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|877.8 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-6-4-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..5553d753e08c0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_6_4_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_6_4_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_6_4_cnndm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_6_4_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_6_4_cnndm_pipeline_en_5.4.2_3.0_1724004848331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_6_4_cnndm_pipeline_en_5.4.2_3.0_1724004848331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_6_4_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_6_4_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_6_4_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|877.8 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-6-4-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_en.md new file mode 100644 index 00000000000000..1c8526a02f3b46 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_alphax T5Transformer from cbharathcs +author: John Snow Labs +name: flan_t5_base_alphax +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_alphax` is a English model originally trained by cbharathcs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_alphax_en_5.4.2_3.0_1723994199360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_alphax_en_5.4.2_3.0_1723994199360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_alphax","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_alphax", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_alphax| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cbharathcs/flan-t5-base-alphaX \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_pipeline_en.md new file mode 100644 index 00000000000000..e2f13129280f38 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_alphax_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_alphax_pipeline pipeline T5Transformer from cbharathcs +author: John Snow Labs +name: flan_t5_base_alphax_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_alphax_pipeline` is a English model originally trained by cbharathcs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_alphax_pipeline_en_5.4.2_3.0_1723994245051.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_alphax_pipeline_en_5.4.2_3.0_1723994245051.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_alphax_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_alphax_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_alphax_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cbharathcs/flan-t5-base-alphaX + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_en.md new file mode 100644 index 00000000000000..a959a14acfa5f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_conditional_translation_me_english_me T5Transformer from zuko2 +author: John Snow Labs +name: flan_t5_base_conditional_translation_me_english_me +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_conditional_translation_me_english_me` is a English model originally trained by zuko2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_conditional_translation_me_english_me_en_5.4.2_3.0_1723995175217.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_conditional_translation_me_english_me_en_5.4.2_3.0_1723995175217.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_conditional_translation_me_english_me","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_conditional_translation_me_english_me", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_conditional_translation_me_english_me| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/zuko2/flan-t5-base-conditional-translation-me-en-me \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_pipeline_en.md new file mode 100644 index 00000000000000..1adf26d6b5f9dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_conditional_translation_me_english_me_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_conditional_translation_me_english_me_pipeline pipeline T5Transformer from zuko2 +author: John Snow Labs +name: flan_t5_base_conditional_translation_me_english_me_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_conditional_translation_me_english_me_pipeline` is a English model originally trained by zuko2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_conditional_translation_me_english_me_pipeline_en_5.4.2_3.0_1723995221477.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_conditional_translation_me_english_me_pipeline_en_5.4.2_3.0_1723995221477.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_conditional_translation_me_english_me_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_conditional_translation_me_english_me_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_conditional_translation_me_english_me_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/zuko2/flan-t5-base-conditional-translation-me-en-me + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_en.md new file mode 100644 index 00000000000000..70d8a46453586c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_esnli T5Transformer from Mitrofazotron +author: John Snow Labs +name: flan_t5_base_esnli +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_esnli` is a English model originally trained by Mitrofazotron. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_esnli_en_5.4.2_3.0_1723994218294.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_esnli_en_5.4.2_3.0_1723994218294.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_esnli","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_esnli", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_esnli| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Mitrofazotron/flan-t5-base-esnli \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_pipeline_en.md new file mode 100644 index 00000000000000..82937c85d42c8c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_esnli_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_esnli_pipeline pipeline T5Transformer from Mitrofazotron +author: John Snow Labs +name: flan_t5_base_esnli_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_esnli_pipeline` is a English model originally trained by Mitrofazotron. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_esnli_pipeline_en_5.4.2_3.0_1723994264122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_esnli_pipeline_en_5.4.2_3.0_1723994264122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_esnli_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_esnli_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_esnli_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Mitrofazotron/flan-t5-base-esnli + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_en.md new file mode 100644 index 00000000000000..90eb1a46f03826 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_extract_issues_and_suggestions_3 T5Transformer from SachinPatel248 +author: John Snow Labs +name: flan_t5_base_extract_issues_and_suggestions_3 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extract_issues_and_suggestions_3` is a English model originally trained by SachinPatel248. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extract_issues_and_suggestions_3_en_5.4.2_3.0_1723965466588.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extract_issues_and_suggestions_3_en_5.4.2_3.0_1723965466588.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_extract_issues_and_suggestions_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_extract_issues_and_suggestions_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extract_issues_and_suggestions_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SachinPatel248/flan-t5-base-extract-issues-and-suggestions_3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_pipeline_en.md new file mode 100644 index 00000000000000..52a3d4117a789d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_extract_issues_and_suggestions_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_extract_issues_and_suggestions_3_pipeline pipeline T5Transformer from SachinPatel248 +author: John Snow Labs +name: flan_t5_base_extract_issues_and_suggestions_3_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extract_issues_and_suggestions_3_pipeline` is a English model originally trained by SachinPatel248. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extract_issues_and_suggestions_3_pipeline_en_5.4.2_3.0_1723965515191.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extract_issues_and_suggestions_3_pipeline_en_5.4.2_3.0_1723965515191.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_extract_issues_and_suggestions_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_extract_issues_and_suggestions_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extract_issues_and_suggestions_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SachinPatel248/flan-t5-base-extract-issues-and-suggestions_3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_en.md new file mode 100644 index 00000000000000..000b3a50e75e13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_10_cogs T5Transformer from pmedepal +author: John Snow Labs +name: flan_t5_base_finetuned_10_cogs +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_10_cogs` is a English model originally trained by pmedepal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_10_cogs_en_5.4.2_3.0_1723964786901.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_10_cogs_en_5.4.2_3.0_1723964786901.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_10_cogs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_10_cogs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_10_cogs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pmedepal/flan-t5-base-finetuned-10-cogs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_pipeline_en.md new file mode 100644 index 00000000000000..f1c2a311d91bdd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_10_cogs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_10_cogs_pipeline pipeline T5Transformer from pmedepal +author: John Snow Labs +name: flan_t5_base_finetuned_10_cogs_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_10_cogs_pipeline` is a English model originally trained by pmedepal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_10_cogs_pipeline_en_5.4.2_3.0_1723964836037.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_10_cogs_pipeline_en_5.4.2_3.0_1723964836037.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_10_cogs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_10_cogs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_10_cogs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pmedepal/flan-t5-base-finetuned-10-cogs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_en.md new file mode 100644 index 00000000000000..fd8a16641c8c79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51 T5Transformer from jerome1519 +author: John Snow Labs +name: flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51` is a English model originally trained by jerome1519. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_en_5.4.2_3.0_1723953733788.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_en_5.4.2_3.0_1723953733788.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jerome1519/flan-t5-base-finetuned-coding_instructions_2023_08_18__07_51 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline_en.md new file mode 100644 index 00000000000000..019c6d4f2dbc6c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline pipeline T5Transformer from jerome1519 +author: John Snow Labs +name: flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline` is a English model originally trained by jerome1519. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline_en_5.4.2_3.0_1723953809542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline_en_5.4.2_3.0_1723953809542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_coding_instructions_2023_08_18__07_51_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jerome1519/flan-t5-base-finetuned-coding_instructions_2023_08_18__07_51 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_en.md new file mode 100644 index 00000000000000..04c150c45c5c2e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_dialougesum_english_spanish T5Transformer from Gowreesh234 +author: John Snow Labs +name: flan_t5_base_finetuned_dialougesum_english_spanish +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_dialougesum_english_spanish` is a English model originally trained by Gowreesh234. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_dialougesum_english_spanish_en_5.4.2_3.0_1723977548982.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_dialougesum_english_spanish_en_5.4.2_3.0_1723977548982.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_dialougesum_english_spanish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_dialougesum_english_spanish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_dialougesum_english_spanish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.4 MB| + +## References + +https://huggingface.co/Gowreesh234/flan-t5-base-finetuned-dialougesum-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_pipeline_en.md new file mode 100644 index 00000000000000..576a911686e460 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_dialougesum_english_spanish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_dialougesum_english_spanish_pipeline pipeline T5Transformer from Gowreesh234 +author: John Snow Labs +name: flan_t5_base_finetuned_dialougesum_english_spanish_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_dialougesum_english_spanish_pipeline` is a English model originally trained by Gowreesh234. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_dialougesum_english_spanish_pipeline_en_5.4.2_3.0_1723977726098.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_dialougesum_english_spanish_pipeline_en_5.4.2_3.0_1723977726098.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_dialougesum_english_spanish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_dialougesum_english_spanish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_dialougesum_english_spanish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.4 MB| + +## References + +https://huggingface.co/Gowreesh234/flan-t5-base-finetuned-dialougesum-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_en.md new file mode 100644 index 00000000000000..33489054559721 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_longlearnrate_mts_dialogue T5Transformer from agnesem +author: John Snow Labs +name: flan_t5_base_finetuned_longlearnrate_mts_dialogue +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_longlearnrate_mts_dialogue` is a English model originally trained by agnesem. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_longlearnrate_mts_dialogue_en_5.4.2_3.0_1723954752281.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_longlearnrate_mts_dialogue_en_5.4.2_3.0_1723954752281.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_longlearnrate_mts_dialogue","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_longlearnrate_mts_dialogue", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_longlearnrate_mts_dialogue| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.3 MB| + +## References + +https://huggingface.co/agnesem/flan_t5_base_finetuned_longLearnRate_MTS_dialogue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline_en.md new file mode 100644 index 00000000000000..9bbc557ac21077 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline pipeline T5Transformer from agnesem +author: John Snow Labs +name: flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline` is a English model originally trained by agnesem. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline_en_5.4.2_3.0_1723954931892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline_en_5.4.2_3.0_1723954931892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_longlearnrate_mts_dialogue_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.3 MB| + +## References + +https://huggingface.co/agnesem/flan_t5_base_finetuned_longLearnRate_MTS_dialogue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_en.md new file mode 100644 index 00000000000000..cb798821c0fadc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_productdomain_instructions T5Transformer from plaskod +author: John Snow Labs +name: flan_t5_base_productdomain_instructions +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_productdomain_instructions` is a English model originally trained by plaskod. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_productdomain_instructions_en_5.4.2_3.0_1724004208654.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_productdomain_instructions_en_5.4.2_3.0_1724004208654.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_productdomain_instructions","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_productdomain_instructions", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_productdomain_instructions| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/plaskod/flan-t5-base-productdomain_instructions \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_pipeline_en.md new file mode 100644 index 00000000000000..bcaaa68b3be5ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_productdomain_instructions_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_productdomain_instructions_pipeline pipeline T5Transformer from plaskod +author: John Snow Labs +name: flan_t5_base_productdomain_instructions_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_productdomain_instructions_pipeline` is a English model originally trained by plaskod. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_productdomain_instructions_pipeline_en_5.4.2_3.0_1724004254492.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_productdomain_instructions_pipeline_en_5.4.2_3.0_1724004254492.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_productdomain_instructions_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_productdomain_instructions_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_productdomain_instructions_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/plaskod/flan-t5-base-productdomain_instructions + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_en.md new file mode 100644 index 00000000000000..829811e1a3af83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_qg_squad_10_v2 T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_base_qg_squad_10_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_qg_squad_10_v2` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_10_v2_en_5.4.2_3.0_1723978680139.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_10_v2_en_5.4.2_3.0_1723978680139.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_qg_squad_10_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_qg_squad_10_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_qg_squad_10_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-base-qg-SQuAD-10-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_pipeline_en.md new file mode 100644 index 00000000000000..624a3da4ce77c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_qg_squad_10_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_qg_squad_10_v2_pipeline pipeline T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_base_qg_squad_10_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_qg_squad_10_v2_pipeline` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_10_v2_pipeline_en_5.4.2_3.0_1723978727710.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_10_v2_pipeline_en_5.4.2_3.0_1723978727710.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_qg_squad_10_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_qg_squad_10_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_qg_squad_10_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-base-qg-SQuAD-10-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_en.md new file mode 100644 index 00000000000000..bc54eda8d3c280 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_reviewss_text_classification T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_reviewss_text_classification +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_reviewss_text_classification` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_reviewss_text_classification_en_5.4.2_3.0_1723943365087.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_reviewss_text_classification_en_5.4.2_3.0_1723943365087.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_reviewss_text_classification","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_reviewss_text_classification", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_reviewss_text_classification| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-reviewss-text-classification \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_pipeline_en.md new file mode 100644 index 00000000000000..a4397ea85f5c0c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_reviewss_text_classification_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_reviewss_text_classification_pipeline pipeline T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_reviewss_text_classification_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_reviewss_text_classification_pipeline` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_reviewss_text_classification_pipeline_en_5.4.2_3.0_1723943416068.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_reviewss_text_classification_pipeline_en_5.4.2_3.0_1723943416068.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_reviewss_text_classification_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_reviewss_text_classification_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_reviewss_text_classification_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-reviewss-text-classification + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_en.md new file mode 100644 index 00000000000000..27409bcb4abba8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_tweet_qg T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_base_tweet_qg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_tweet_qg` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qg_en_5.4.2_3.0_1723947744305.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qg_en_5.4.2_3.0_1723947744305.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_tweet_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_tweet_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_tweet_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-base-tweet-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_pipeline_en.md new file mode 100644 index 00000000000000..3d9ea36b5d13a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_base_tweet_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_tweet_qg_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_base_tweet_qg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_tweet_qg_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qg_pipeline_en_5.4.2_3.0_1723947796297.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qg_pipeline_en_5.4.2_3.0_1723947796297.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_tweet_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_tweet_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_tweet_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-base-tweet-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_en.md new file mode 100644 index 00000000000000..2f2ce47b0df382 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_continuity_rephrase_v1_0 T5Transformer from magnifi +author: John Snow Labs +name: flan_t5_continuity_rephrase_v1_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_continuity_rephrase_v1_0` is a English model originally trained by magnifi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_continuity_rephrase_v1_0_en_5.4.2_3.0_1723993493169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_continuity_rephrase_v1_0_en_5.4.2_3.0_1723993493169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_continuity_rephrase_v1_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_continuity_rephrase_v1_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_continuity_rephrase_v1_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/magnifi/flan-t5-continuity-rephrase-v1.0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_pipeline_en.md new file mode 100644 index 00000000000000..b8595cbe21ea3d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_continuity_rephrase_v1_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_continuity_rephrase_v1_0_pipeline pipeline T5Transformer from magnifi +author: John Snow Labs +name: flan_t5_continuity_rephrase_v1_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_continuity_rephrase_v1_0_pipeline` is a English model originally trained by magnifi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_continuity_rephrase_v1_0_pipeline_en_5.4.2_3.0_1723993542468.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_continuity_rephrase_v1_0_pipeline_en_5.4.2_3.0_1723993542468.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_continuity_rephrase_v1_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_continuity_rephrase_v1_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_continuity_rephrase_v1_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/magnifi/flan-t5-continuity-rephrase-v1.0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_en.md new file mode 100644 index 00000000000000..5e07caf15f0622 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_glue_finetuning_lr8e_4 T5Transformer from liuyanchen1015 +author: John Snow Labs +name: flan_t5_glue_finetuning_lr8e_4 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_glue_finetuning_lr8e_4` is a English model originally trained by liuyanchen1015. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_glue_finetuning_lr8e_4_en_5.4.2_3.0_1723949881537.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_glue_finetuning_lr8e_4_en_5.4.2_3.0_1723949881537.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_glue_finetuning_lr8e_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_glue_finetuning_lr8e_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_glue_finetuning_lr8e_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/liuyanchen1015/FLAN-T5_GLUE_finetuning_lr8e-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_pipeline_en.md new file mode 100644 index 00000000000000..2ea2e6d6285f4d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_glue_finetuning_lr8e_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_glue_finetuning_lr8e_4_pipeline pipeline T5Transformer from liuyanchen1015 +author: John Snow Labs +name: flan_t5_glue_finetuning_lr8e_4_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_glue_finetuning_lr8e_4_pipeline` is a English model originally trained by liuyanchen1015. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_glue_finetuning_lr8e_4_pipeline_en_5.4.2_3.0_1723949930749.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_glue_finetuning_lr8e_4_pipeline_en_5.4.2_3.0_1723949930749.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_glue_finetuning_lr8e_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_glue_finetuning_lr8e_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_glue_finetuning_lr8e_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/liuyanchen1015/FLAN-T5_GLUE_finetuning_lr8e-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_en.md new file mode 100644 index 00000000000000..b9f56c44ebe1f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep10 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep10 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep10` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep10_en_5.4.2_3.0_1723941383823.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep10_en_5.4.2_3.0_1723941383823.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline_en.md new file mode 100644 index 00000000000000..5f1368562cdcdb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline_en_5.4.2_3.0_1723941559063.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline_en_5.4.2_3.0_1723941559063.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_en.md new file mode 100644 index 00000000000000..1fcb649ea10612 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_en_5.4.2_3.0_1723962526762.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_en_5.4.2_3.0_1723962526762.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-ep15-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..9ef909efc102e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline_en_5.4.2_3.0_1723962675655.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline_en_5.4.2_3.0_1723962675655.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_ep15_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-ep15-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_en.md new file mode 100644 index 00000000000000..1fbb8a510c5339 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_fs0_2 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_fs0_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_fs0_2` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_2_en_5.4.2_3.0_1723941381418.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_2_en_5.4.2_3.0_1723941381418.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_fs0_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_fs0_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_fs0_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_fs0.2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline_en.md new file mode 100644 index 00000000000000..222821a3b78172 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline_en_5.4.2_3.0_1723941544884.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline_en_5.4.2_3.0_1723941544884.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_fs0_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_fs0.2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_en.md new file mode 100644 index 00000000000000..122f5d157ae721 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_extract_issues_and_suggestions T5Transformer from SachinPatel248 +author: John Snow Labs +name: flan_t5_large_extract_issues_and_suggestions +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extract_issues_and_suggestions` is a English model originally trained by SachinPatel248. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extract_issues_and_suggestions_en_5.4.2_3.0_1723964891524.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extract_issues_and_suggestions_en_5.4.2_3.0_1723964891524.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_extract_issues_and_suggestions","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_extract_issues_and_suggestions", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extract_issues_and_suggestions| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/SachinPatel248/flan-t5-large-extract-issues-and-suggestions \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_pipeline_en.md new file mode 100644 index 00000000000000..4522c4605c213e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extract_issues_and_suggestions_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_extract_issues_and_suggestions_pipeline pipeline T5Transformer from SachinPatel248 +author: John Snow Labs +name: flan_t5_large_extract_issues_and_suggestions_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extract_issues_and_suggestions_pipeline` is a English model originally trained by SachinPatel248. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extract_issues_and_suggestions_pipeline_en_5.4.2_3.0_1723965036582.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extract_issues_and_suggestions_pipeline_en_5.4.2_3.0_1723965036582.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_extract_issues_and_suggestions_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_extract_issues_and_suggestions_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extract_issues_and_suggestions_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/SachinPatel248/flan-t5-large-extract-issues-and-suggestions + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_en.md new file mode 100644 index 00000000000000..a5248a0dac9f42 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_extraction_all_cnn_4000_ep25_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_all_cnn_4000_ep25_nonstop +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_all_cnn_4000_ep25_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_en_5.4.2_3.0_1723992914576.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_en_5.4.2_3.0_1723992914576.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_extraction_all_cnn_4000_ep25_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_extraction_all_cnn_4000_ep25_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_all_cnn_4000_ep25_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-all-cnn_4000-ep25-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..0492f81890d4eb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline_en_5.4.2_3.0_1723993053900.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline_en_5.4.2_3.0_1723993053900.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_all_cnn_4000_ep25_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-all-cnn_4000-ep25-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20_en.md new file mode 100644 index 00000000000000..f4d7ac4f3ad205 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20_en_5.4.2_3.0_1723956609944.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20_en_5.4.2_3.0_1723956609944.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_cnndm_4000_all_hint_hit_ep20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-cnndm_4000-all-hint_hit-ep20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_en.md new file mode 100644 index 00000000000000..3e1d41db2eae80 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_google T5Transformer from google +author: John Snow Labs +name: flan_t5_large_google +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_google` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_google_en_5.4.2_3.0_1724011719259.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_google_en_5.4.2_3.0_1724011719259.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_google","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_google", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_google| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/google/flan-t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_pipeline_en.md new file mode 100644 index 00000000000000..19f8657c9b1041 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_google_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_google_pipeline pipeline T5Transformer from google +author: John Snow Labs +name: flan_t5_large_google_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_google_pipeline` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_google_pipeline_en_5.4.2_3.0_1724011865606.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_google_pipeline_en_5.4.2_3.0_1724011865606.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_google_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_google_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_google_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/google/flan-t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_en.md new file mode 100644 index 00000000000000..ce34a997433e25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_noun_completion_english T5Transformer from Kartikey95 +author: John Snow Labs +name: flan_t5_large_noun_completion_english +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_noun_completion_english` is a English model originally trained by Kartikey95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_noun_completion_english_en_5.4.2_3.0_1723963209228.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_noun_completion_english_en_5.4.2_3.0_1723963209228.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_noun_completion_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_noun_completion_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_noun_completion_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Kartikey95/flan-t5-large-noun-completion-en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_pipeline_en.md new file mode 100644 index 00000000000000..4c3dc55d201c4f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_noun_completion_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_noun_completion_english_pipeline pipeline T5Transformer from Kartikey95 +author: John Snow Labs +name: flan_t5_large_noun_completion_english_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_noun_completion_english_pipeline` is a English model originally trained by Kartikey95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_noun_completion_english_pipeline_en_5.4.2_3.0_1723963346689.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_noun_completion_english_pipeline_en_5.4.2_3.0_1723963346689.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_noun_completion_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_noun_completion_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_noun_completion_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Kartikey95/flan-t5-large-noun-completion-en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_en.md new file mode 100644 index 00000000000000..24f3dd16642b00 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_question_answering_finetuining T5Transformer from lvcalucioli +author: John Snow Labs +name: flan_t5_large_question_answering_finetuining +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_question_answering_finetuining` is a English model originally trained by lvcalucioli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_question_answering_finetuining_en_5.4.2_3.0_1723939795517.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_question_answering_finetuining_en_5.4.2_3.0_1723939795517.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_question_answering_finetuining","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_question_answering_finetuining", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_question_answering_finetuining| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/lvcalucioli/flan-t5-large_question_answering_finetuining \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_pipeline_en.md new file mode 100644 index 00000000000000..b5e2bd83a682ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_large_question_answering_finetuining_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_question_answering_finetuining_pipeline pipeline T5Transformer from lvcalucioli +author: John Snow Labs +name: flan_t5_large_question_answering_finetuining_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_question_answering_finetuining_pipeline` is a English model originally trained by lvcalucioli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_question_answering_finetuining_pipeline_en_5.4.2_3.0_1723939935640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_question_answering_finetuining_pipeline_en_5.4.2_3.0_1723939935640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_question_answering_finetuining_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_question_answering_finetuining_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_question_answering_finetuining_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/lvcalucioli/flan-t5-large_question_answering_finetuining + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_en.md new file mode 100644 index 00000000000000..6c86da594647b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_qg_tarek_test_squad T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_qg_tarek_test_squad +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_qg_tarek_test_squad` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_qg_tarek_test_squad_en_5.4.2_3.0_1724003504026.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_qg_tarek_test_squad_en_5.4.2_3.0_1724003504026.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_qg_tarek_test_squad","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_qg_tarek_test_squad", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_qg_tarek_test_squad| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-qg-tarek-test-SQUAD \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_pipeline_en.md new file mode 100644 index 00000000000000..00fd8a34010e91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_qg_tarek_test_squad_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_qg_tarek_test_squad_pipeline pipeline T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_qg_tarek_test_squad_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_qg_tarek_test_squad_pipeline` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_qg_tarek_test_squad_pipeline_en_5.4.2_3.0_1724003550332.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_qg_tarek_test_squad_pipeline_en_5.4.2_3.0_1724003550332.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_qg_tarek_test_squad_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_qg_tarek_test_squad_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_qg_tarek_test_squad_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-qg-tarek-test-SQUAD + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_en.md new file mode 100644 index 00000000000000..bd74d63696380e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_retacred_all_kongo_w_context_small_finetuned T5Transformer from kinshuk-h +author: John Snow Labs +name: flan_t5_retacred_all_kongo_w_context_small_finetuned +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_retacred_all_kongo_w_context_small_finetuned` is a English model originally trained by kinshuk-h. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_all_kongo_w_context_small_finetuned_en_5.4.2_3.0_1723978752507.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_all_kongo_w_context_small_finetuned_en_5.4.2_3.0_1723978752507.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_retacred_all_kongo_w_context_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_retacred_all_kongo_w_context_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_retacred_all_kongo_w_context_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/kinshuk-h/flan-t5-retacred-all-kg-w-context-small-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..f44bbe6887cd5a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline pipeline T5Transformer from kinshuk-h +author: John Snow Labs +name: flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline` is a English model originally trained by kinshuk-h. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline_en_5.4.2_3.0_1723978769400.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline_en_5.4.2_3.0_1723978769400.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_retacred_all_kongo_w_context_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/kinshuk-h/flan-t5-retacred-all-kg-w-context-small-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_en.md new file mode 100644 index 00000000000000..c5ceb31d6cd319 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_6_1_xsum T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_1_xsum +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_1_xsum` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_1_xsum_en_5.4.2_3.0_1724002582759.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_1_xsum_en_5.4.2_3.0_1724002582759.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_6_1_xsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_6_1_xsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_1_xsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|267.2 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-1-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_pipeline_en.md new file mode 100644 index 00000000000000..5dee8f5d558a6e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_1_xsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_6_1_xsum_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_1_xsum_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_1_xsum_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_1_xsum_pipeline_en_5.4.2_3.0_1724002594845.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_1_xsum_pipeline_en_5.4.2_3.0_1724002594845.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_6_1_xsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_6_1_xsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_1_xsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|267.3 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-1-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_en.md new file mode 100644 index 00000000000000..7266eb250330cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_6_2_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_2_cnndm +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_2_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_2_cnndm_en_5.4.2_3.0_1723994544616.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_2_cnndm_en_5.4.2_3.0_1723994544616.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_6_2_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_6_2_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_2_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|279.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-2-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..ea28d235a9cd99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_6_2_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_6_2_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_2_cnndm_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_2_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_2_cnndm_pipeline_en_5.4.2_3.0_1723994557878.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_2_cnndm_pipeline_en_5.4.2_3.0_1723994557878.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_6_2_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_6_2_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_2_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|279.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-2-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_en.md new file mode 100644 index 00000000000000..ffbf49af5f3f48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_dialogsum_test_500 T5Transformer from MoralesTP +author: John Snow Labs +name: flan_t5_small_dialogsum_test_500 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_dialogsum_test_500` is a English model originally trained by MoralesTP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_dialogsum_test_500_en_5.4.2_3.0_1723997372649.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_dialogsum_test_500_en_5.4.2_3.0_1723997372649.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_dialogsum_test_500","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_dialogsum_test_500", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_dialogsum_test_500| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/MoralesTP/flan-t5-small-dialogsum-test-500 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_pipeline_en.md new file mode 100644 index 00000000000000..606673ff468047 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_dialogsum_test_500_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_dialogsum_test_500_pipeline pipeline T5Transformer from MoralesTP +author: John Snow Labs +name: flan_t5_small_dialogsum_test_500_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_dialogsum_test_500_pipeline` is a English model originally trained by MoralesTP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_dialogsum_test_500_pipeline_en_5.4.2_3.0_1723997388376.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_dialogsum_test_500_pipeline_en_5.4.2_3.0_1723997388376.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_dialogsum_test_500_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_dialogsum_test_500_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_dialogsum_test_500_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/MoralesTP/flan-t5-small-dialogsum-test-500 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_en.md new file mode 100644 index 00000000000000..e08d503ace2fab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_entity T5Transformer from gangiswag +author: John Snow Labs +name: flan_t5_small_entity +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_entity` is a English model originally trained by gangiswag. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_entity_en_5.4.2_3.0_1724001530638.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_entity_en_5.4.2_3.0_1724001530638.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_entity","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_entity", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_entity| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/gangiswag/flan_t5_small_entity \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_pipeline_en.md new file mode 100644 index 00000000000000..dab9c2c5f4bfc3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_entity_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_entity_pipeline pipeline T5Transformer from gangiswag +author: John Snow Labs +name: flan_t5_small_entity_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_entity_pipeline` is a English model originally trained by gangiswag. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_entity_pipeline_en_5.4.2_3.0_1724001668526.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_entity_pipeline_en_5.4.2_3.0_1724001668526.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_entity_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_entity_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_entity_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/gangiswag/flan_t5_small_entity + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_en.md new file mode 100644 index 00000000000000..30cba7fe46fdc8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_5 T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_5` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_5_en_5.4.2_3.0_1724001808398.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_5_en_5.4.2_3.0_1724001808398.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_pipeline_en.md new file mode 100644 index 00000000000000..05077170a8ffc0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_finetuned_coqa_v0_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_5_pipeline pipeline T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_5_pipeline` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_5_pipeline_en_5.4.2_3.0_1724001826752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_5_pipeline_en_5.4.2_3.0_1724001826752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_en.md new file mode 100644 index 00000000000000..64b587d5b32efe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_gsm8k T5Transformer from fiveflow +author: John Snow Labs +name: flan_t5_small_gsm8k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_gsm8k` is a English model originally trained by fiveflow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_gsm8k_en_5.4.2_3.0_1723997211497.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_gsm8k_en_5.4.2_3.0_1723997211497.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_gsm8k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_gsm8k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_gsm8k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/fiveflow/flan-t5-small-gsm8k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_pipeline_en.md new file mode 100644 index 00000000000000..de5c0574f218e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_gsm8k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_gsm8k_pipeline pipeline T5Transformer from fiveflow +author: John Snow Labs +name: flan_t5_small_gsm8k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_gsm8k_pipeline` is a English model originally trained by fiveflow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_gsm8k_pipeline_en_5.4.2_3.0_1723997228496.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_gsm8k_pipeline_en_5.4.2_3.0_1723997228496.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_gsm8k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_gsm8k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_gsm8k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/fiveflow/flan-t5-small-gsm8k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_en.md new file mode 100644 index 00000000000000..5d8c817fc7108e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_qa_9 T5Transformer from badokorach +author: John Snow Labs +name: flan_t5_small_qa_9 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_qa_9` is a English model originally trained by badokorach. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_qa_9_en_5.4.2_3.0_1723976734211.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_qa_9_en_5.4.2_3.0_1723976734211.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_qa_9","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_qa_9", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_qa_9| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/badokorach/flan-t5-small-qa-9 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_pipeline_en.md new file mode 100644 index 00000000000000..b9b3463777215c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_qa_9_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_qa_9_pipeline pipeline T5Transformer from badokorach +author: John Snow Labs +name: flan_t5_small_qa_9_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_qa_9_pipeline` is a English model originally trained by badokorach. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_qa_9_pipeline_en_5.4.2_3.0_1723976783306.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_qa_9_pipeline_en_5.4.2_3.0_1723976783306.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_qa_9_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_qa_9_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_qa_9_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/badokorach/flan-t5-small-qa-9 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_en.md new file mode 100644 index 00000000000000..1c0b5f389d0a04 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_samsum_emilijatr T5Transformer from EmilijaTR +author: John Snow Labs +name: flan_t5_small_samsum_emilijatr +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_emilijatr` is a English model originally trained by EmilijaTR. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_emilijatr_en_5.4.2_3.0_1723983930540.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_emilijatr_en_5.4.2_3.0_1723983930540.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_samsum_emilijatr","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_samsum_emilijatr", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_emilijatr| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/EmilijaTR/flan-t5-small-samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_pipeline_en.md new file mode 100644 index 00000000000000..acd194b9d104b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_samsum_emilijatr_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_samsum_emilijatr_pipeline pipeline T5Transformer from EmilijaTR +author: John Snow Labs +name: flan_t5_small_samsum_emilijatr_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_emilijatr_pipeline` is a English model originally trained by EmilijaTR. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_emilijatr_pipeline_en_5.4.2_3.0_1723983946776.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_emilijatr_pipeline_en_5.4.2_3.0_1723983946776.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_samsum_emilijatr_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_samsum_emilijatr_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_emilijatr_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/EmilijaTR/flan-t5-small-samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_en.md new file mode 100644 index 00000000000000..adfb9ec16a6643 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_section_32_qna T5Transformer from Kelum +author: John Snow Labs +name: flan_t5_small_section_32_qna +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_section_32_qna` is a English model originally trained by Kelum. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_section_32_qna_en_5.4.2_3.0_1723943670068.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_section_32_qna_en_5.4.2_3.0_1723943670068.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_section_32_qna","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_section_32_qna", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_section_32_qna| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/Kelum/Flan_T5_small_section_32_QnA \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_pipeline_en.md new file mode 100644 index 00000000000000..11160881e306ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_section_32_qna_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_section_32_qna_pipeline pipeline T5Transformer from Kelum +author: John Snow Labs +name: flan_t5_small_section_32_qna_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_section_32_qna_pipeline` is a English model originally trained by Kelum. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_section_32_qna_pipeline_en_5.4.2_3.0_1723943687409.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_section_32_qna_pipeline_en_5.4.2_3.0_1723943687409.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_section_32_qna_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_section_32_qna_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_section_32_qna_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/Kelum/Flan_T5_small_section_32_QnA + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_en.md new file mode 100644 index 00000000000000..0f82e025c62c32 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_tweet_ner7 T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_ner7 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_ner7` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_ner7_en_5.4.2_3.0_1724001230287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_ner7_en_5.4.2_3.0_1724001230287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_tweet_ner7","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_tweet_ner7", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_ner7| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-ner7 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_pipeline_en.md new file mode 100644 index 00000000000000..ea8f4e9affb5f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_ner7_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_tweet_ner7_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_ner7_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_ner7_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_ner7_pipeline_en_5.4.2_3.0_1724001246236.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_ner7_pipeline_en_5.4.2_3.0_1724001246236.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_tweet_ner7_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_tweet_ner7_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_ner7_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-ner7 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_en.md new file mode 100644 index 00000000000000..51dfde886187ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_tweet_topic T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_topic +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_topic` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_topic_en_5.4.2_3.0_1724021421830.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_topic_en_5.4.2_3.0_1724021421830.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_tweet_topic","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_tweet_topic", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_topic| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-topic \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_pipeline_en.md new file mode 100644 index 00000000000000..b431d3125c2d48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_small_tweet_topic_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_tweet_topic_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_topic_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_topic_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_topic_pipeline_en_5.4.2_3.0_1724021443598.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_topic_pipeline_en_5.4.2_3.0_1724021443598.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_tweet_topic_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_tweet_topic_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_topic_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-topic + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_en.md new file mode 100644 index 00000000000000..fbedeeda5cd62e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_summerize_legal_doc_padded T5Transformer from shlomik +author: John Snow Labs +name: flan_t5_summerize_legal_doc_padded +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_summerize_legal_doc_padded` is a English model originally trained by shlomik. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_summerize_legal_doc_padded_en_5.4.2_3.0_1724007997246.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_summerize_legal_doc_padded_en_5.4.2_3.0_1724007997246.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_summerize_legal_doc_padded","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_summerize_legal_doc_padded", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_summerize_legal_doc_padded| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/shlomik/flan-T5-summerize-legal-doc-padded \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_pipeline_en.md new file mode 100644 index 00000000000000..be7a770c339b4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flan_t5_summerize_legal_doc_padded_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_summerize_legal_doc_padded_pipeline pipeline T5Transformer from shlomik +author: John Snow Labs +name: flan_t5_summerize_legal_doc_padded_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_summerize_legal_doc_padded_pipeline` is a English model originally trained by shlomik. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_summerize_legal_doc_padded_pipeline_en_5.4.2_3.0_1724008125002.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_summerize_legal_doc_padded_pipeline_en_5.4.2_3.0_1724008125002.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_summerize_legal_doc_padded_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_summerize_legal_doc_padded_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_summerize_legal_doc_padded_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/shlomik/flan-T5-summerize-legal-doc-padded + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_en.md new file mode 100644 index 00000000000000..b08b2f2c4c5d09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_base_mcfatbeard T5Transformer from mcfatbeard +author: John Snow Labs +name: flant5_base_mcfatbeard +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_base_mcfatbeard` is a English model originally trained by mcfatbeard. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_base_mcfatbeard_en_5.4.2_3.0_1724006000685.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_base_mcfatbeard_en_5.4.2_3.0_1724006000685.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_base_mcfatbeard","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_base_mcfatbeard", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_base_mcfatbeard| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mcfatbeard/flant5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_pipeline_en.md new file mode 100644 index 00000000000000..5b672d522f19ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_base_mcfatbeard_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_base_mcfatbeard_pipeline pipeline T5Transformer from mcfatbeard +author: John Snow Labs +name: flant5_base_mcfatbeard_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_base_mcfatbeard_pipeline` is a English model originally trained by mcfatbeard. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_base_mcfatbeard_pipeline_en_5.4.2_3.0_1724006045631.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_base_mcfatbeard_pipeline_en_5.4.2_3.0_1724006045631.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_base_mcfatbeard_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_base_mcfatbeard_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_base_mcfatbeard_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mcfatbeard/flant5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_en.md new file mode 100644 index 00000000000000..a09b46c1c959cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_dolly_qna_prompt T5Transformer from perfectlybaked +author: John Snow Labs +name: flant5_dolly_qna_prompt +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_dolly_qna_prompt` is a English model originally trained by perfectlybaked. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_dolly_qna_prompt_en_5.4.2_3.0_1723994927077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_dolly_qna_prompt_en_5.4.2_3.0_1723994927077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_dolly_qna_prompt","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_dolly_qna_prompt", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_dolly_qna_prompt| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/perfectlybaked/flant5-dolly-QnA-prompt \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_pipeline_en.md new file mode 100644 index 00000000000000..199fa4745e2208 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_dolly_qna_prompt_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_dolly_qna_prompt_pipeline pipeline T5Transformer from perfectlybaked +author: John Snow Labs +name: flant5_dolly_qna_prompt_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_dolly_qna_prompt_pipeline` is a English model originally trained by perfectlybaked. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_dolly_qna_prompt_pipeline_en_5.4.2_3.0_1723994973291.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_dolly_qna_prompt_pipeline_en_5.4.2_3.0_1723994973291.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_dolly_qna_prompt_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_dolly_qna_prompt_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_dolly_qna_prompt_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/perfectlybaked/flant5-dolly-QnA-prompt + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_en.md new file mode 100644 index 00000000000000..0fc9bf901cc298 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_large_aio_a1_b2_c3_d4_archana T5Transformer from a1-b2-c3-d4-archana +author: John Snow Labs +name: flant5_large_aio_a1_b2_c3_d4_archana +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_large_aio_a1_b2_c3_d4_archana` is a English model originally trained by a1-b2-c3-d4-archana. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_large_aio_a1_b2_c3_d4_archana_en_5.4.2_3.0_1724010570150.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_large_aio_a1_b2_c3_d4_archana_en_5.4.2_3.0_1724010570150.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_large_aio_a1_b2_c3_d4_archana","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_large_aio_a1_b2_c3_d4_archana", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_large_aio_a1_b2_c3_d4_archana| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/a1-b2-c3-d4-archana/flant5-large-aio \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_pipeline_en.md new file mode 100644 index 00000000000000..e75eb1dc08317d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-flant5_large_aio_a1_b2_c3_d4_archana_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_large_aio_a1_b2_c3_d4_archana_pipeline pipeline T5Transformer from a1-b2-c3-d4-archana +author: John Snow Labs +name: flant5_large_aio_a1_b2_c3_d4_archana_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_large_aio_a1_b2_c3_d4_archana_pipeline` is a English model originally trained by a1-b2-c3-d4-archana. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_large_aio_a1_b2_c3_d4_archana_pipeline_en_5.4.2_3.0_1724010703261.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_large_aio_a1_b2_c3_d4_archana_pipeline_en_5.4.2_3.0_1724010703261.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_large_aio_a1_b2_c3_d4_archana_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_large_aio_a1_b2_c3_d4_archana_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_large_aio_a1_b2_c3_d4_archana_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/a1-b2-c3-d4-archana/flant5-large-aio + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_en.md b/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_en.md new file mode 100644 index 00000000000000..6027669a9a55a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_aeroqa_1hop_top5_colbert_20 T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_top5_colbert_20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_top5_colbert_20` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_top5_colbert_20_en_5.4.2_3.0_1723999329603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_top5_colbert_20_en_5.4.2_3.0_1723999329603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_aeroqa_1hop_top5_colbert_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_aeroqa_1hop_top5_colbert_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_top5_colbert_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_top5_COLBERT_20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_pipeline_en.md new file mode 100644 index 00000000000000..5309f57d079bbd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ft_aeroqa_1hop_top5_colbert_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ft_aeroqa_1hop_top5_colbert_20_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_top5_colbert_20_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_top5_colbert_20_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_top5_colbert_20_pipeline_en_5.4.2_3.0_1723999462239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_top5_colbert_20_pipeline_en_5.4.2_3.0_1723999462239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ft_aeroqa_1hop_top5_colbert_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ft_aeroqa_1hop_top5_colbert_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_top5_colbert_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_top5_COLBERT_20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ft_metaqa_1hop_kongo_unverb_20_20_skill_en.md b/docs/_posts/ahmedlone127/2024-08-18-ft_metaqa_1hop_kongo_unverb_20_20_skill_en.md new file mode 100644 index 00000000000000..f8853f1d8d201d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ft_metaqa_1hop_kongo_unverb_20_20_skill_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_metaqa_1hop_kongo_unverb_20_20_skill T5Transformer from sakharamg +author: John Snow Labs +name: ft_metaqa_1hop_kongo_unverb_20_20_skill +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_metaqa_1hop_kongo_unverb_20_20_skill` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_metaqa_1hop_kongo_unverb_20_20_skill_en_5.4.2_3.0_1723957156786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_metaqa_1hop_kongo_unverb_20_20_skill_en_5.4.2_3.0_1723957156786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_metaqa_1hop_kongo_unverb_20_20_skill","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_metaqa_1hop_kongo_unverb_20_20_skill", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_metaqa_1hop_kongo_unverb_20_20_skill| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_metaqa_1hop_kg_unverb_20_20_SKILL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_en.md b/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_en.md new file mode 100644 index 00000000000000..cd3b5128e92cfb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English gner_t5_base T5Transformer from dyyyyyyyy +author: John Snow Labs +name: gner_t5_base +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`gner_t5_base` is a English model originally trained by dyyyyyyyy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/gner_t5_base_en_5.4.2_3.0_1724022574404.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/gner_t5_base_en_5.4.2_3.0_1724022574404.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("gner_t5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("gner_t5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|gner_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/dyyyyyyyy/GNER-T5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_pipeline_en.md new file mode 100644 index 00000000000000..5004dc8f6383b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-gner_t5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English gner_t5_base_pipeline pipeline T5Transformer from dyyyyyyyy +author: John Snow Labs +name: gner_t5_base_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`gner_t5_base_pipeline` is a English model originally trained by dyyyyyyyy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/gner_t5_base_pipeline_en_5.4.2_3.0_1724022620068.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/gner_t5_base_pipeline_en_5.4.2_3.0_1724022620068.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("gner_t5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("gner_t5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|gner_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/dyyyyyyyy/GNER-T5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_en.md b/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_en.md new file mode 100644 index 00000000000000..e4fb4c0b309604 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English google_flan_t5_base_alpaca T5Transformer from reasonwang +author: John Snow Labs +name: google_flan_t5_base_alpaca +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_flan_t5_base_alpaca` is a English model originally trained by reasonwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_flan_t5_base_alpaca_en_5.4.2_3.0_1724023860784.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_flan_t5_base_alpaca_en_5.4.2_3.0_1724023860784.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("google_flan_t5_base_alpaca","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("google_flan_t5_base_alpaca", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_flan_t5_base_alpaca| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/reasonwang/google-flan-t5-base-alpaca \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_pipeline_en.md new file mode 100644 index 00000000000000..e54ead203bf246 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-google_flan_t5_base_alpaca_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English google_flan_t5_base_alpaca_pipeline pipeline T5Transformer from reasonwang +author: John Snow Labs +name: google_flan_t5_base_alpaca_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_flan_t5_base_alpaca_pipeline` is a English model originally trained by reasonwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_flan_t5_base_alpaca_pipeline_en_5.4.2_3.0_1724023906690.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_flan_t5_base_alpaca_pipeline_en_5.4.2_3.0_1724023906690.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("google_flan_t5_base_alpaca_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("google_flan_t5_base_alpaca_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_flan_t5_base_alpaca_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/reasonwang/google-flan-t5-base-alpaca + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_en.md b/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_en.md new file mode 100644 index 00000000000000..bb95606dcfbf1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English grammar_correction_model T5Transformer from hassaanik +author: John Snow Labs +name: grammar_correction_model +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`grammar_correction_model` is a English model originally trained by hassaanik. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/grammar_correction_model_en_5.4.2_3.0_1724013761254.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/grammar_correction_model_en_5.4.2_3.0_1724013761254.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("grammar_correction_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("grammar_correction_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|grammar_correction_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|985.1 MB| + +## References + +https://huggingface.co/hassaanik/grammar-correction-model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_pipeline_en.md new file mode 100644 index 00000000000000..19ea81223428f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-grammar_correction_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English grammar_correction_model_pipeline pipeline T5Transformer from hassaanik +author: John Snow Labs +name: grammar_correction_model_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`grammar_correction_model_pipeline` is a English model originally trained by hassaanik. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/grammar_correction_model_pipeline_en_5.4.2_3.0_1724013812982.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/grammar_correction_model_pipeline_en_5.4.2_3.0_1724013812982.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("grammar_correction_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("grammar_correction_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|grammar_correction_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|985.1 MB| + +## References + +https://huggingface.co/hassaanik/grammar-correction-model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_en.md new file mode 100644 index 00000000000000..2b65a718bc3947 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English green_words_flan_t5_small T5Transformer from gizmo-ai +author: John Snow Labs +name: green_words_flan_t5_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`green_words_flan_t5_small` is a English model originally trained by gizmo-ai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/green_words_flan_t5_small_en_5.4.2_3.0_1723978629270.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/green_words_flan_t5_small_en_5.4.2_3.0_1723978629270.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("green_words_flan_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("green_words_flan_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|green_words_flan_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|178.6 MB| + +## References + +https://huggingface.co/gizmo-ai/green-words-flan-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..7f326f5dbba91e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-green_words_flan_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English green_words_flan_t5_small_pipeline pipeline T5Transformer from gizmo-ai +author: John Snow Labs +name: green_words_flan_t5_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`green_words_flan_t5_small_pipeline` is a English model originally trained by gizmo-ai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/green_words_flan_t5_small_pipeline_en_5.4.2_3.0_1723978690396.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/green_words_flan_t5_small_pipeline_en_5.4.2_3.0_1723978690396.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("green_words_flan_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("green_words_flan_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|green_words_flan_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|178.6 MB| + +## References + +https://huggingface.co/gizmo-ai/green-words-flan-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-happy_gec_en.md b/docs/_posts/ahmedlone127/2024-08-18-happy_gec_en.md new file mode 100644 index 00000000000000..b2fafd15b8cd8f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-happy_gec_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English happy_gec T5Transformer from audribean +author: John Snow Labs +name: happy_gec +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_gec` is a English model originally trained by audribean. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_gec_en_5.4.2_3.0_1723944582571.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_gec_en_5.4.2_3.0_1723944582571.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("happy_gec","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("happy_gec", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_gec| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|971.5 MB| + +## References + +https://huggingface.co/audribean/happy-gec \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-happy_gec_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-happy_gec_pipeline_en.md new file mode 100644 index 00000000000000..7e854c2cda48f1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-happy_gec_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English happy_gec_pipeline pipeline T5Transformer from audribean +author: John Snow Labs +name: happy_gec_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_gec_pipeline` is a English model originally trained by audribean. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_gec_pipeline_en_5.4.2_3.0_1723944642648.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_gec_pipeline_en_5.4.2_3.0_1723944642648.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("happy_gec_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("happy_gec_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_gec_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|971.5 MB| + +## References + +https://huggingface.co/audribean/happy-gec + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_en.md new file mode 100644 index 00000000000000..cdec950a082b99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_ep_v2 T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_ep_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_ep_v2` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v2_en_5.4.2_3.0_1723947590744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v2_en_5.4.2_3.0_1723947590744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_ep_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_ep_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_ep_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-ep-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_pipeline_en.md new file mode 100644 index 00000000000000..c1520cd9e2c69c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-happy_transformer_t5_base_grammar_correction_ep_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_ep_v2_pipeline pipeline T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_ep_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_ep_v2_pipeline` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v2_pipeline_en_5.4.2_3.0_1723947640791.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v2_pipeline_en_5.4.2_3.0_1723947640791.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("happy_transformer_t5_base_grammar_correction_ep_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("happy_transformer_t5_base_grammar_correction_ep_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_ep_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-ep-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_en.md b/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_en.md new file mode 100644 index 00000000000000..4ab2d7b38c1408 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English indot5_small_qg T5Transformer from widyanto +author: John Snow Labs +name: indot5_small_qg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`indot5_small_qg` is a English model originally trained by widyanto. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/indot5_small_qg_en_5.4.2_3.0_1723943046008.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/indot5_small_qg_en_5.4.2_3.0_1723943046008.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("indot5_small_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("indot5_small_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|indot5_small_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|294.5 MB| + +## References + +https://huggingface.co/widyanto/IndoT5-small-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_pipeline_en.md new file mode 100644 index 00000000000000..07ee8c59474d00 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-indot5_small_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English indot5_small_qg_pipeline pipeline T5Transformer from widyanto +author: John Snow Labs +name: indot5_small_qg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`indot5_small_qg_pipeline` is a English model originally trained by widyanto. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/indot5_small_qg_pipeline_en_5.4.2_3.0_1723943077177.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/indot5_small_qg_pipeline_en_5.4.2_3.0_1723943077177.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("indot5_small_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("indot5_small_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|indot5_small_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|294.5 MB| + +## References + +https://huggingface.co/widyanto/IndoT5-small-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_en.md b/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_en.md new file mode 100644 index 00000000000000..462225bec7f72a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English italian_spanish_80k T5Transformer from Bistolero +author: John Snow Labs +name: italian_spanish_80k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_spanish_80k` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_spanish_80k_en_5.4.2_3.0_1723998849885.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_spanish_80k_en_5.4.2_3.0_1723998849885.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("italian_spanish_80k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("italian_spanish_80k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_spanish_80k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/it_es_80k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_pipeline_en.md new file mode 100644 index 00000000000000..e297f32f913140 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-italian_spanish_80k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English italian_spanish_80k_pipeline pipeline T5Transformer from Bistolero +author: John Snow Labs +name: italian_spanish_80k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_spanish_80k_pipeline` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_spanish_80k_pipeline_en_5.4.2_3.0_1723999002121.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_spanish_80k_pipeline_en_5.4.2_3.0_1723999002121.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("italian_spanish_80k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("italian_spanish_80k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_spanish_80k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/it_es_80k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_en.md b/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_en.md new file mode 100644 index 00000000000000..16e7e0a86bad8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English italian_train_all T5Transformer from Bistolero +author: John Snow Labs +name: italian_train_all +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_train_all` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_train_all_en_5.4.2_3.0_1724004785258.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_train_all_en_5.4.2_3.0_1724004785258.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("italian_train_all","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("italian_train_all", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_train_all| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Bistolero/it_train_all \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_pipeline_en.md new file mode 100644 index 00000000000000..f27087a75eb985 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-italian_train_all_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English italian_train_all_pipeline pipeline T5Transformer from Bistolero +author: John Snow Labs +name: italian_train_all_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`italian_train_all_pipeline` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/italian_train_all_pipeline_en_5.4.2_3.0_1724004951421.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/italian_train_all_pipeline_en_5.4.2_3.0_1724004951421.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("italian_train_all_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("italian_train_all_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|italian_train_all_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Bistolero/it_train_all + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-jdpoc_en.md b/docs/_posts/ahmedlone127/2024-08-18-jdpoc_en.md new file mode 100644 index 00000000000000..e617be3f8d7790 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-jdpoc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English jdpoc T5Transformer from Mawiah +author: John Snow Labs +name: jdpoc +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`jdpoc` is a English model originally trained by Mawiah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/jdpoc_en_5.4.2_3.0_1723999344487.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/jdpoc_en_5.4.2_3.0_1723999344487.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("jdpoc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("jdpoc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|jdpoc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/Mawiah/JDPOC \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-jdpoc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-jdpoc_pipeline_en.md new file mode 100644 index 00000000000000..a8c1736ffc6d6b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-jdpoc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English jdpoc_pipeline pipeline T5Transformer from Mawiah +author: John Snow Labs +name: jdpoc_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`jdpoc_pipeline` is a English model originally trained by Mawiah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/jdpoc_pipeline_en_5.4.2_3.0_1723999361641.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/jdpoc_pipeline_en_5.4.2_3.0_1723999361641.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("jdpoc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("jdpoc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|jdpoc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/Mawiah/JDPOC + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_en.md b/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_en.md new file mode 100644 index 00000000000000..b6d1cedb65864f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kb13_t5_base_finetuned_english_tonga_tonga_islands_regex T5Transformer from rymaju +author: John Snow Labs +name: kb13_t5_base_finetuned_english_tonga_tonga_islands_regex +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kb13_t5_base_finetuned_english_tonga_tonga_islands_regex` is a English model originally trained by rymaju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_en_5.4.2_3.0_1724022122773.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_en_5.4.2_3.0_1724022122773.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kb13_t5_base_finetuned_english_tonga_tonga_islands_regex","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kb13_t5_base_finetuned_english_tonga_tonga_islands_regex", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kb13_t5_base_finetuned_english_tonga_tonga_islands_regex| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/rymaju/KB13-t5-base-finetuned-en-to-regex \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline_en.md new file mode 100644 index 00000000000000..341b22de3012e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline pipeline T5Transformer from rymaju +author: John Snow Labs +name: kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline` is a English model originally trained by rymaju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline_en_5.4.2_3.0_1724022170380.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline_en_5.4.2_3.0_1724022170380.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kb13_t5_base_finetuned_english_tonga_tonga_islands_regex_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/rymaju/KB13-t5-base-finetuned-en-to-regex + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_en.md b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_en.md new file mode 100644 index 00000000000000..74a5f5204e505d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english T5Transformer from JIEUN21 +author: John Snow Labs +name: ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english` is a English model originally trained by JIEUN21. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_en_5.4.2_3.0_1723993378179.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_en_5.4.2_3.0_1723993378179.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|663.5 MB| + +## References + +https://huggingface.co/JIEUN21/ke-t5-finetuned-kde4-ko-to-en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline_en.md new file mode 100644 index 00000000000000..87b7a67f215f09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline pipeline T5Transformer from JIEUN21 +author: John Snow Labs +name: ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline` is a English model originally trained by JIEUN21. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline_en_5.4.2_3.0_1723993594692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline_en_5.4.2_3.0_1723993594692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_finetuned_kde4_korean_tonga_tonga_islands_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|663.5 MB| + +## References + +https://huggingface.co/JIEUN21/ke-t5-finetuned-kde4-ko-to-en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_en.md b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_en.md new file mode 100644 index 00000000000000..58fb644a6175c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ke_t5_small_finetuned_amazon_english_spanish T5Transformer from alphahg +author: John Snow Labs +name: ke_t5_small_finetuned_amazon_english_spanish +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_small_finetuned_amazon_english_spanish` is a English model originally trained by alphahg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_small_finetuned_amazon_english_spanish_en_5.4.2_3.0_1723985730907.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_small_finetuned_amazon_english_spanish_en_5.4.2_3.0_1723985730907.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ke_t5_small_finetuned_amazon_english_spanish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ke_t5_small_finetuned_amazon_english_spanish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_small_finetuned_amazon_english_spanish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|493.6 MB| + +## References + +https://huggingface.co/alphahg/ke-t5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_pipeline_en.md new file mode 100644 index 00000000000000..763ea1a94e219d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ke_t5_small_finetuned_amazon_english_spanish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ke_t5_small_finetuned_amazon_english_spanish_pipeline pipeline T5Transformer from alphahg +author: John Snow Labs +name: ke_t5_small_finetuned_amazon_english_spanish_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_small_finetuned_amazon_english_spanish_pipeline` is a English model originally trained by alphahg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_small_finetuned_amazon_english_spanish_pipeline_en_5.4.2_3.0_1723985759859.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_small_finetuned_amazon_english_spanish_pipeline_en_5.4.2_3.0_1723985759859.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ke_t5_small_finetuned_amazon_english_spanish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ke_t5_small_finetuned_amazon_english_spanish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_small_finetuned_amazon_english_spanish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|493.6 MB| + +## References + +https://huggingface.co/alphahg/ke-t5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_en.md new file mode 100644 index 00000000000000..f31dbb7db308cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kingify_2way_t5_large_v1_1 T5Transformer from swcrazyfan +author: John Snow Labs +name: kingify_2way_t5_large_v1_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kingify_2way_t5_large_v1_1` is a English model originally trained by swcrazyfan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kingify_2way_t5_large_v1_1_en_5.4.2_3.0_1724023205035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kingify_2way_t5_large_v1_1_en_5.4.2_3.0_1724023205035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kingify_2way_t5_large_v1_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kingify_2way_t5_large_v1_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kingify_2way_t5_large_v1_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/swcrazyfan/Kingify-2Way-T5-Large-v1_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_pipeline_en.md new file mode 100644 index 00000000000000..26ad8658cfb6d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kingify_2way_t5_large_v1_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kingify_2way_t5_large_v1_1_pipeline pipeline T5Transformer from swcrazyfan +author: John Snow Labs +name: kingify_2way_t5_large_v1_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kingify_2way_t5_large_v1_1_pipeline` is a English model originally trained by swcrazyfan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kingify_2way_t5_large_v1_1_pipeline_en_5.4.2_3.0_1724023351527.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kingify_2way_t5_large_v1_1_pipeline_en_5.4.2_3.0_1724023351527.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kingify_2way_t5_large_v1_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kingify_2way_t5_large_v1_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kingify_2way_t5_large_v1_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/swcrazyfan/Kingify-2Way-T5-Large-v1_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_sapol_v5_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_sapol_v5_en.md new file mode 100644 index 00000000000000..2b7c1e5e1a8053 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_sapol_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_sapol_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_sapol_v5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_sapol_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sapol_v5_en_5.4.2_3.0_1723949802871.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sapol_v5_en_5.4.2_3.0_1723949802871.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_sapol_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_sapol_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_sapol_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SAPOL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_en.md new file mode 100644 index 00000000000000..3df79917b8b662 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_spoal_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_spoal_v5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_spoal_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spoal_v5_en_5.4.2_3.0_1723964980392.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spoal_v5_en_5.4.2_3.0_1723964980392.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_spoal_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_spoal_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_spoal_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SPOAL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_pipeline_en.md new file mode 100644 index 00000000000000..7a97f2fb3e1c5e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_spoal_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_spoal_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_spoal_v5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_spoal_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spoal_v5_pipeline_en_5.4.2_3.0_1723965155482.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_spoal_v5_pipeline_en_5.4.2_3.0_1723965155482.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_spoal_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_spoal_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_spoal_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SPOAL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_apsol_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_apsol_en.md new file mode 100644 index 00000000000000..88cde0c48bdd4d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_apsol_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_apsol T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_apsol +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_apsol` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_apsol_en_5.4.2_3.0_1723985769266.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_apsol_en_5.4.2_3.0_1723985769266.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_apsol","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_apsol", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_apsol| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_APSOL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_en.md new file mode 100644 index 00000000000000..8ea89835f2cf7a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_poasl_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_poasl_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_poasl_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v2_en_5.4.2_3.0_1723984326366.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v2_en_5.4.2_3.0_1723984326366.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_poasl_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_poasl_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_poasl_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POASL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_pipeline_en.md new file mode 100644 index 00000000000000..d8948ea538b690 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_poasl_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_poasl_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_poasl_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_poasl_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v2_pipeline_en_5.4.2_3.0_1723984499823.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v2_pipeline_en_5.4.2_3.0_1723984499823.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_poasl_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_poasl_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_poasl_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POASL_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_en.md new file mode 100644 index 00000000000000..3dd980c6df70de --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v2_en_5.4.2_3.0_1724003408411.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v2_en_5.4.2_3.0_1724003408411.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_pipeline_en.md new file mode 100644 index 00000000000000..0653aa59750cb5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-kltn_coqe_vit5_total_sapol_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v2_pipeline_en_5.4.2_3.0_1724003573248.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_v2_pipeline_en_5.4.2_3.0_1724003573248.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_sapol_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_sapol_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_en.md b/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_en.md new file mode 100644 index 00000000000000..ed83241ba34e1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lamini_flan_t5_248m_orca_12_5k_4bit T5Transformer from marcchew +author: John Snow Labs +name: lamini_flan_t5_248m_orca_12_5k_4bit +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lamini_flan_t5_248m_orca_12_5k_4bit` is a English model originally trained by marcchew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_248m_orca_12_5k_4bit_en_5.4.2_3.0_1723991741007.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_248m_orca_12_5k_4bit_en_5.4.2_3.0_1723991741007.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lamini_flan_t5_248m_orca_12_5k_4bit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lamini_flan_t5_248m_orca_12_5k_4bit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lamini_flan_t5_248m_orca_12_5k_4bit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|558.0 MB| + +## References + +https://huggingface.co/marcchew/LaMini-Flan-T5-248M-Orca-12.5K-4bit \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_pipeline_en.md new file mode 100644 index 00000000000000..bd0581904f74d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lamini_flan_t5_248m_orca_12_5k_4bit_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lamini_flan_t5_248m_orca_12_5k_4bit_pipeline pipeline T5Transformer from marcchew +author: John Snow Labs +name: lamini_flan_t5_248m_orca_12_5k_4bit_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lamini_flan_t5_248m_orca_12_5k_4bit_pipeline` is a English model originally trained by marcchew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_248m_orca_12_5k_4bit_pipeline_en_5.4.2_3.0_1723991876549.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_248m_orca_12_5k_4bit_pipeline_en_5.4.2_3.0_1723991876549.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lamini_flan_t5_248m_orca_12_5k_4bit_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lamini_flan_t5_248m_orca_12_5k_4bit_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lamini_flan_t5_248m_orca_12_5k_4bit_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|558.0 MB| + +## References + +https://huggingface.co/marcchew/LaMini-Flan-T5-248M-Orca-12.5K-4bit + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_en.md new file mode 100644 index 00000000000000..2a066862b13247 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_flan_t5_chunked T5Transformer from Ashreen +author: John Snow Labs +name: legal_flan_t5_chunked +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_flan_t5_chunked` is a English model originally trained by Ashreen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_flan_t5_chunked_en_5.4.2_3.0_1723993707929.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_flan_t5_chunked_en_5.4.2_3.0_1723993707929.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_flan_t5_chunked","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_flan_t5_chunked", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_flan_t5_chunked| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Ashreen/legal-flan-t5-chunked \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_pipeline_en.md new file mode 100644 index 00000000000000..e58644549b9852 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_flan_t5_chunked_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_flan_t5_chunked_pipeline pipeline T5Transformer from Ashreen +author: John Snow Labs +name: legal_flan_t5_chunked_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_flan_t5_chunked_pipeline` is a English model originally trained by Ashreen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_flan_t5_chunked_pipeline_en_5.4.2_3.0_1723993842209.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_flan_t5_chunked_pipeline_en_5.4.2_3.0_1723993842209.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_flan_t5_chunked_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_flan_t5_chunked_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_flan_t5_chunked_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Ashreen/legal-flan-t5-chunked + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_en.md new file mode 100644 index 00000000000000..67e7302bd46136 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_large_hybrid T5Transformer from Ashreen +author: John Snow Labs +name: legal_t5_large_hybrid +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_large_hybrid` is a English model originally trained by Ashreen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_large_hybrid_en_5.4.2_3.0_1723991385730.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_large_hybrid_en_5.4.2_3.0_1723991385730.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_large_hybrid","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_large_hybrid", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_large_hybrid| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Ashreen/legal-t5-large-hybrid \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_pipeline_en.md new file mode 100644 index 00000000000000..69f07af2519701 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_large_hybrid_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_large_hybrid_pipeline pipeline T5Transformer from Ashreen +author: John Snow Labs +name: legal_t5_large_hybrid_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_large_hybrid_pipeline` is a English model originally trained by Ashreen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_large_hybrid_pipeline_en_5.4.2_3.0_1723991523836.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_large_hybrid_pipeline_en_5.4.2_3.0_1723991523836.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_large_hybrid_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_large_hybrid_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_large_hybrid_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Ashreen/legal-t5-large-hybrid + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_en.md new file mode 100644 index 00000000000000..436c5f508e752a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_cls_finetuned_french T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_cls_finetuned_french +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_cls_finetuned_french` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_finetuned_french_en_5.4.2_3.0_1724019948346.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_finetuned_french_en_5.4.2_3.0_1724019948346.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_cls_finetuned_french","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_cls_finetuned_french", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_cls_finetuned_french| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_cls_finetuned_fr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_pipeline_en.md new file mode 100644 index 00000000000000..052ddc13b654a2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_cls_finetuned_french_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_cls_finetuned_french_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_cls_finetuned_french_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_cls_finetuned_french_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_finetuned_french_pipeline_en_5.4.2_3.0_1724020006295.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_finetuned_french_pipeline_en_5.4.2_3.0_1724020006295.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_cls_finetuned_french_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_cls_finetuned_french_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_cls_finetuned_french_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.2 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_cls_finetuned_fr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_en.md new file mode 100644 index 00000000000000..d2508058729d02 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_finetuned_summ_french T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_finetuned_summ_french +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_finetuned_summ_french` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_finetuned_summ_french_en_5.4.2_3.0_1723995818985.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_finetuned_summ_french_en_5.4.2_3.0_1723995818985.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_finetuned_summ_french","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_finetuned_summ_french", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_finetuned_summ_french| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.2 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_finetuned_summ_fr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_pipeline_en.md new file mode 100644 index 00000000000000..8f8f6e3c5d572b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_finetuned_summ_french_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_finetuned_summ_french_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_finetuned_summ_french_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_finetuned_summ_french_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_finetuned_summ_french_pipeline_en_5.4.2_3.0_1723995876022.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_finetuned_summ_french_pipeline_en_5.4.2_3.0_1723995876022.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_finetuned_summ_french_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_finetuned_summ_french_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_finetuned_summ_french_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.2 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_finetuned_summ_fr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_en.md new file mode 100644 index 00000000000000..89a8b54c61e143 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_multitask_swedish_czech T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_swedish_czech +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_swedish_czech` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_swedish_czech_en_5.4.2_3.0_1724024445280.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_swedish_czech_en_5.4.2_3.0_1724024445280.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_multitask_swedish_czech","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_multitask_swedish_czech", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_swedish_czech| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_sv_cs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_pipeline_en.md new file mode 100644 index 00000000000000..8bce5f0b426eb8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_multitask_swedish_czech_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_multitask_swedish_czech_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_swedish_czech_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_swedish_czech_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_swedish_czech_pipeline_en_5.4.2_3.0_1724024504359.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_swedish_czech_pipeline_en_5.4.2_3.0_1724024504359.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_multitask_swedish_czech_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_multitask_swedish_czech_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_swedish_czech_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_sv_cs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_en.md new file mode 100644 index 00000000000000..0609384b533476 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_czech T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_czech +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_czech` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_czech_en_5.4.2_3.0_1724007432089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_czech_en_5.4.2_3.0_1724007432089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_czech","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_czech", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_czech| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|180.0 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_cs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_pipeline_en.md new file mode 100644 index 00000000000000..7066678740398e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_czech_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_czech_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_czech_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_czech_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_czech_pipeline_en_5.4.2_3.0_1724007489828.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_czech_pipeline_en_5.4.2_3.0_1724007489828.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_italian_czech_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_italian_czech_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_czech_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|180.0 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_cs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_en.md new file mode 100644 index 00000000000000..7404d45577ec8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_swedish T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_swedish +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_swedish` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_swedish_en_5.4.2_3.0_1723989121196.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_swedish_en_5.4.2_3.0_1723989121196.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_swedish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_swedish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_swedish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_sv \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_pipeline_en.md new file mode 100644 index 00000000000000..9fa97bcfbd78c7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-legal_t5_small_trans_italian_swedish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_swedish_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_swedish_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_swedish_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_swedish_pipeline_en_5.4.2_3.0_1723989178439.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_swedish_pipeline_en_5.4.2_3.0_1723989178439.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_italian_swedish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_italian_swedish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_swedish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_sv + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_ko.md b/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_ko.md new file mode 100644 index 00000000000000..b26f2eab27ae87 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_ko.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Korean long_ke_t5_base_summarization T5Transformer from KETI-AIR-Downstream +author: John Snow Labs +name: long_ke_t5_base_summarization +date: 2024-08-18 +tags: [ko, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_ke_t5_base_summarization` is a Korean model originally trained by KETI-AIR-Downstream. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_ke_t5_base_summarization_ko_5.4.2_3.0_1724016567817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_ke_t5_base_summarization_ko_5.4.2_3.0_1724016567817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("long_ke_t5_base_summarization","ko") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("long_ke_t5_base_summarization", "ko") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_ke_t5_base_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ko| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KETI-AIR-Downstream/long-ke-t5-base-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_pipeline_ko.md b/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_pipeline_ko.md new file mode 100644 index 00000000000000..fef44ab04676f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-long_ke_t5_base_summarization_pipeline_ko.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Korean long_ke_t5_base_summarization_pipeline pipeline T5Transformer from KETI-AIR-Downstream +author: John Snow Labs +name: long_ke_t5_base_summarization_pipeline +date: 2024-08-18 +tags: [ko, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_ke_t5_base_summarization_pipeline` is a Korean model originally trained by KETI-AIR-Downstream. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_ke_t5_base_summarization_pipeline_ko_5.4.2_3.0_1724016629050.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_ke_t5_base_summarization_pipeline_ko_5.4.2_3.0_1724016629050.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("long_ke_t5_base_summarization_pipeline", lang = "ko") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("long_ke_t5_base_summarization_pipeline", lang = "ko") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_ke_t5_base_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ko| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KETI-AIR-Downstream/long-ke-t5-base-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_en.md b/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_en.md new file mode 100644 index 00000000000000..36d32353b8c636 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English long_t5_local_base_finetuned_justification_v10 T5Transformer from satyanshu404 +author: John Snow Labs +name: long_t5_local_base_finetuned_justification_v10 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5_local_base_finetuned_justification_v10` is a English model originally trained by satyanshu404. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5_local_base_finetuned_justification_v10_en_5.4.2_3.0_1723990940401.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5_local_base_finetuned_justification_v10_en_5.4.2_3.0_1723990940401.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("long_t5_local_base_finetuned_justification_v10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("long_t5_local_base_finetuned_justification_v10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5_local_base_finetuned_justification_v10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|743.6 MB| + +## References + +https://huggingface.co/satyanshu404/long-t5-local-base-finetuned-justification-v10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_pipeline_en.md new file mode 100644 index 00000000000000..c3338a514b337d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-long_t5_local_base_finetuned_justification_v10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English long_t5_local_base_finetuned_justification_v10_pipeline pipeline T5Transformer from satyanshu404 +author: John Snow Labs +name: long_t5_local_base_finetuned_justification_v10_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5_local_base_finetuned_justification_v10_pipeline` is a English model originally trained by satyanshu404. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5_local_base_finetuned_justification_v10_pipeline_en_5.4.2_3.0_1723990975253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5_local_base_finetuned_justification_v10_pipeline_en_5.4.2_3.0_1723990975253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("long_t5_local_base_finetuned_justification_v10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("long_t5_local_base_finetuned_justification_v10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5_local_base_finetuned_justification_v10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|743.6 MB| + +## References + +https://huggingface.co/satyanshu404/long-t5-local-base-finetuned-justification-v10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_en.md b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_en.md new file mode 100644 index 00000000000000..d894abb37149d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lora_flan_t5_chuvash_chatbot T5Transformer from saadrasheeddev +author: John Snow Labs +name: lora_flan_t5_chuvash_chatbot +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lora_flan_t5_chuvash_chatbot` is a English model originally trained by saadrasheeddev. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lora_flan_t5_chuvash_chatbot_en_5.4.2_3.0_1724006797656.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lora_flan_t5_chuvash_chatbot_en_5.4.2_3.0_1724006797656.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lora_flan_t5_chuvash_chatbot","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lora_flan_t5_chuvash_chatbot", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lora_flan_t5_chuvash_chatbot| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/saadrasheeddev/lora-flan-t5-cv-chatbot \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_pipeline_en.md new file mode 100644 index 00000000000000..5d6daab1fe42d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_chuvash_chatbot_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lora_flan_t5_chuvash_chatbot_pipeline pipeline T5Transformer from saadrasheeddev +author: John Snow Labs +name: lora_flan_t5_chuvash_chatbot_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lora_flan_t5_chuvash_chatbot_pipeline` is a English model originally trained by saadrasheeddev. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lora_flan_t5_chuvash_chatbot_pipeline_en_5.4.2_3.0_1724006930623.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lora_flan_t5_chuvash_chatbot_pipeline_en_5.4.2_3.0_1724006930623.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lora_flan_t5_chuvash_chatbot_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lora_flan_t5_chuvash_chatbot_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lora_flan_t5_chuvash_chatbot_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/saadrasheeddev/lora-flan-t5-cv-chatbot + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_en.md b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_en.md new file mode 100644 index 00000000000000..18245c1ccdaa65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lora_flan_t5_large_chat_saadrasheeddev T5Transformer from saadrasheeddev +author: John Snow Labs +name: lora_flan_t5_large_chat_saadrasheeddev +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lora_flan_t5_large_chat_saadrasheeddev` is a English model originally trained by saadrasheeddev. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lora_flan_t5_large_chat_saadrasheeddev_en_5.4.2_3.0_1724012028461.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lora_flan_t5_large_chat_saadrasheeddev_en_5.4.2_3.0_1724012028461.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lora_flan_t5_large_chat_saadrasheeddev","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lora_flan_t5_large_chat_saadrasheeddev", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lora_flan_t5_large_chat_saadrasheeddev| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/saadrasheeddev/lora-flan-t5-large-chat \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_pipeline_en.md new file mode 100644 index 00000000000000..2511c80eaee9b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-lora_flan_t5_large_chat_saadrasheeddev_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lora_flan_t5_large_chat_saadrasheeddev_pipeline pipeline T5Transformer from saadrasheeddev +author: John Snow Labs +name: lora_flan_t5_large_chat_saadrasheeddev_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lora_flan_t5_large_chat_saadrasheeddev_pipeline` is a English model originally trained by saadrasheeddev. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lora_flan_t5_large_chat_saadrasheeddev_pipeline_en_5.4.2_3.0_1724012155409.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lora_flan_t5_large_chat_saadrasheeddev_pipeline_en_5.4.2_3.0_1724012155409.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lora_flan_t5_large_chat_saadrasheeddev_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lora_flan_t5_large_chat_saadrasheeddev_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lora_flan_t5_large_chat_saadrasheeddev_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/saadrasheeddev/lora-flan-t5-large-chat + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-low5_en.md b/docs/_posts/ahmedlone127/2024-08-18-low5_en.md new file mode 100644 index 00000000000000..93c7cf1d1e3e1f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-low5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English low5 T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: low5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`low5` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/low5_en_5.4.2_3.0_1723947865516.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/low5_en_5.4.2_3.0_1723947865516.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("low5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("low5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|low5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|984.4 MB| + +## References + +https://huggingface.co/AliGhiasvand86/low5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-low5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-low5_pipeline_en.md new file mode 100644 index 00000000000000..bcaa1b594167d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-low5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English low5_pipeline pipeline T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: low5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`low5_pipeline` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/low5_pipeline_en_5.4.2_3.0_1723947922450.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/low5_pipeline_en_5.4.2_3.0_1723947922450.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("low5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("low5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|low5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|984.4 MB| + +## References + +https://huggingface.co/AliGhiasvand86/low5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_en.md new file mode 100644 index 00000000000000..f9bdfe4f108745 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_1511_v8 T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_1511_v8 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_1511_v8` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_1511_v8_en_5.4.2_3.0_1723951631354.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_1511_v8_en_5.4.2_3.0_1723951631354.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_1511_v8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_1511_v8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_1511_v8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_1511_v8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_pipeline_en.md new file mode 100644 index 00000000000000..51f6e9e93f3e5f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_1511_v8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_1511_v8_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_1511_v8_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_1511_v8_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_1511_v8_pipeline_en_5.4.2_3.0_1723951793366.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_1511_v8_pipeline_en_5.4.2_3.0_1723951793366.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_1511_v8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_1511_v8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_1511_v8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_1511_v8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_en.md new file mode 100644 index 00000000000000..c68c7f9af26c5e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_2311_retrain_v19_iwt T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_2311_retrain_v19_iwt +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_2311_retrain_v19_iwt` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_2311_retrain_v19_iwt_en_5.4.2_3.0_1723940540513.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_2311_retrain_v19_iwt_en_5.4.2_3.0_1723940540513.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_2311_retrain_v19_iwt","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_2311_retrain_v19_iwt", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_2311_retrain_v19_iwt| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_2311_retrain_v19_iwt \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_pipeline_en.md new file mode 100644 index 00000000000000..bb1e2c64e87724 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_2311_retrain_v19_iwt_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_2311_retrain_v19_iwt_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_2311_retrain_v19_iwt_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_2311_retrain_v19_iwt_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_2311_retrain_v19_iwt_pipeline_en_5.4.2_3.0_1723940696411.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_2311_retrain_v19_iwt_pipeline_en_5.4.2_3.0_1723940696411.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_2311_retrain_v19_iwt_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_2311_retrain_v19_iwt_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_2311_retrain_v19_iwt_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_2311_retrain_v19_iwt + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_en.md new file mode 100644 index 00000000000000..c891d22f933cfe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_base_boun_first_2012_v2 T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_first_2012_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_first_2012_v2` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_v2_en_5.4.2_3.0_1723997567586.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_v2_en_5.4.2_3.0_1723997567586.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_base_boun_first_2012_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_base_boun_first_2012_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_first_2012_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_first_2012_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_pipeline_en.md new file mode 100644 index 00000000000000..e0af92c741abf1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-md_mt5_base_boun_first_2012_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_base_boun_first_2012_v2_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_first_2012_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_first_2012_v2_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_v2_pipeline_en_5.4.2_3.0_1723997714930.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_first_2012_v2_pipeline_en_5.4.2_3.0_1723997714930.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_base_boun_first_2012_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_base_boun_first_2012_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_first_2012_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_first_2012_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_en.md b/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_en.md new file mode 100644 index 00000000000000..5a101f3bcde1b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English medical_summarizer T5Transformer from minz27 +author: John Snow Labs +name: medical_summarizer +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`medical_summarizer` is a English model originally trained by minz27. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/medical_summarizer_en_5.4.2_3.0_1723990041900.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/medical_summarizer_en_5.4.2_3.0_1723990041900.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("medical_summarizer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("medical_summarizer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|medical_summarizer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.9 MB| + +## References + +https://huggingface.co/minz27/medical_summarizer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_pipeline_en.md new file mode 100644 index 00000000000000..3eab69efa96b71 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-medical_summarizer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English medical_summarizer_pipeline pipeline T5Transformer from minz27 +author: John Snow Labs +name: medical_summarizer_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`medical_summarizer_pipeline` is a English model originally trained by minz27. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/medical_summarizer_pipeline_en_5.4.2_3.0_1723990060235.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/medical_summarizer_pipeline_en_5.4.2_3.0_1723990060235.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("medical_summarizer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("medical_summarizer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|medical_summarizer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.9 MB| + +## References + +https://huggingface.co/minz27/medical_summarizer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_en.md b/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_en.md new file mode 100644 index 00000000000000..6116f25360ecb2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model_t5_pramilamanick T5Transformer from Pramilamanick +author: John Snow Labs +name: model_t5_pramilamanick +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_t5_pramilamanick` is a English model originally trained by Pramilamanick. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_t5_pramilamanick_en_5.4.2_3.0_1724006563178.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_t5_pramilamanick_en_5.4.2_3.0_1724006563178.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model_t5_pramilamanick","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model_t5_pramilamanick", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_t5_pramilamanick| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|995.4 MB| + +## References + +https://huggingface.co/Pramilamanick/model_T5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_pipeline_en.md new file mode 100644 index 00000000000000..c5cb14659cbd64 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-model_t5_pramilamanick_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model_t5_pramilamanick_pipeline pipeline T5Transformer from Pramilamanick +author: John Snow Labs +name: model_t5_pramilamanick_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_t5_pramilamanick_pipeline` is a English model originally trained by Pramilamanick. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_t5_pramilamanick_pipeline_en_5.4.2_3.0_1724006612104.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_t5_pramilamanick_pipeline_en_5.4.2_3.0_1724006612104.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model_t5_pramilamanick_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model_t5_pramilamanick_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_t5_pramilamanick_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|995.4 MB| + +## References + +https://huggingface.co/Pramilamanick/model_T5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_en.md new file mode 100644 index 00000000000000..840af36b643bb4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model_text_summarization T5Transformer from nguyethoa +author: John Snow Labs +name: model_text_summarization +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_text_summarization` is a English model originally trained by nguyethoa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_text_summarization_en_5.4.2_3.0_1723966706108.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_text_summarization_en_5.4.2_3.0_1723966706108.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model_text_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model_text_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_text_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/nguyethoa/model_text_summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_pipeline_en.md new file mode 100644 index 00000000000000..6d8eb9719a1faf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-model_text_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model_text_summarization_pipeline pipeline T5Transformer from nguyethoa +author: John Snow Labs +name: model_text_summarization_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_text_summarization_pipeline` is a English model originally trained by nguyethoa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_text_summarization_pipeline_en_5.4.2_3.0_1723966755776.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_text_summarization_pipeline_en_5.4.2_3.0_1723966755776.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model_text_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model_text_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_text_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/nguyethoa/model_text_summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_en.md b/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_en.md new file mode 100644 index 00000000000000..ee0bee5e217b45 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English molt5_small_smiles2caption_lm24 T5Transformer from cnedwards +author: John Snow Labs +name: molt5_small_smiles2caption_lm24 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`molt5_small_smiles2caption_lm24` is a English model originally trained by cnedwards. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/molt5_small_smiles2caption_lm24_en_5.4.2_3.0_1723991005850.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/molt5_small_smiles2caption_lm24_en_5.4.2_3.0_1723991005850.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("molt5_small_smiles2caption_lm24","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("molt5_small_smiles2caption_lm24", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|molt5_small_smiles2caption_lm24| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/cnedwards/molt5-small-smiles2caption-LM24 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_pipeline_en.md new file mode 100644 index 00000000000000..3dc0e51c5abd51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-molt5_small_smiles2caption_lm24_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English molt5_small_smiles2caption_lm24_pipeline pipeline T5Transformer from cnedwards +author: John Snow Labs +name: molt5_small_smiles2caption_lm24_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`molt5_small_smiles2caption_lm24_pipeline` is a English model originally trained by cnedwards. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/molt5_small_smiles2caption_lm24_pipeline_en_5.4.2_3.0_1723991022209.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/molt5_small_smiles2caption_lm24_pipeline_en_5.4.2_3.0_1723991022209.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("molt5_small_smiles2caption_lm24_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("molt5_small_smiles2caption_lm24_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|molt5_small_smiles2caption_lm24_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/cnedwards/molt5-small-smiles2caption-LM24 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_en.md new file mode 100644 index 00000000000000..e0254ac031e495 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mpf_google_t5_samsum_3_epochs_finetuned T5Transformer from StDestiny +author: John Snow Labs +name: mpf_google_t5_samsum_3_epochs_finetuned +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mpf_google_t5_samsum_3_epochs_finetuned` is a English model originally trained by StDestiny. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mpf_google_t5_samsum_3_epochs_finetuned_en_5.4.2_3.0_1724004228474.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mpf_google_t5_samsum_3_epochs_finetuned_en_5.4.2_3.0_1724004228474.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mpf_google_t5_samsum_3_epochs_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mpf_google_t5_samsum_3_epochs_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mpf_google_t5_samsum_3_epochs_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|989.7 MB| + +## References + +https://huggingface.co/StDestiny/MPF-google-t5-samsum-3-epochs-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..db19f54140bca3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mpf_google_t5_samsum_3_epochs_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mpf_google_t5_samsum_3_epochs_finetuned_pipeline pipeline T5Transformer from StDestiny +author: John Snow Labs +name: mpf_google_t5_samsum_3_epochs_finetuned_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mpf_google_t5_samsum_3_epochs_finetuned_pipeline` is a English model originally trained by StDestiny. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mpf_google_t5_samsum_3_epochs_finetuned_pipeline_en_5.4.2_3.0_1724004278413.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mpf_google_t5_samsum_3_epochs_finetuned_pipeline_en_5.4.2_3.0_1724004278413.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mpf_google_t5_samsum_3_epochs_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mpf_google_t5_samsum_3_epochs_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mpf_google_t5_samsum_3_epochs_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|989.7 MB| + +## References + +https://huggingface.co/StDestiny/MPF-google-t5-samsum-3-epochs-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_en.md new file mode 100644 index 00000000000000..16f8ad5e9ed3f5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mrc_v1 T5Transformer from hoangphu7122002ai +author: John Snow Labs +name: mrc_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mrc_v1` is a English model originally trained by hoangphu7122002ai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mrc_v1_en_5.4.2_3.0_1723955195278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mrc_v1_en_5.4.2_3.0_1723955195278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mrc_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mrc_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mrc_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hoangphu7122002ai/MRC_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_pipeline_en.md new file mode 100644 index 00000000000000..51b986fe6da509 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mrc_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mrc_v1_pipeline pipeline T5Transformer from hoangphu7122002ai +author: John Snow Labs +name: mrc_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mrc_v1_pipeline` is a English model originally trained by hoangphu7122002ai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mrc_v1_pipeline_en_5.4.2_3.0_1723955249536.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mrc_v1_pipeline_en_5.4.2_3.0_1723955249536.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mrc_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mrc_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mrc_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hoangphu7122002ai/MRC_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-msmarco_japanese_mt5_base_v1_ja.md b/docs/_posts/ahmedlone127/2024-08-18-msmarco_japanese_mt5_base_v1_ja.md new file mode 100644 index 00000000000000..d7547afa98c5c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-msmarco_japanese_mt5_base_v1_ja.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Japanese msmarco_japanese_mt5_base_v1 T5Transformer from doc2query +author: John Snow Labs +name: msmarco_japanese_mt5_base_v1 +date: 2024-08-18 +tags: [ja, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`msmarco_japanese_mt5_base_v1` is a Japanese model originally trained by doc2query. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/msmarco_japanese_mt5_base_v1_ja_5.4.2_3.0_1724020176463.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/msmarco_japanese_mt5_base_v1_ja_5.4.2_3.0_1724020176463.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("msmarco_japanese_mt5_base_v1","ja") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("msmarco_japanese_mt5_base_v1", "ja") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|msmarco_japanese_mt5_base_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|2.5 GB| + +## References + +https://huggingface.co/doc2query/msmarco-japanese-mt5-base-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_en.md new file mode 100644 index 00000000000000..df7b264d9f4d53 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_20000 T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_20000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_20000` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_20000_en_5.4.2_3.0_1723998485127.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_20000_en_5.4.2_3.0_1723998485127.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_20000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_20000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_20000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_20000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_pipeline_en.md new file mode 100644 index 00000000000000..dc9c2db8a969bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_20000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_20000_pipeline pipeline T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_20000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_20000_pipeline` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_20000_pipeline_en_5.4.2_3.0_1723998683663.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_20000_pipeline_en_5.4.2_3.0_1723998683663.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_20000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_20000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_20000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_20000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_en.md new file mode 100644 index 00000000000000..1972f647924347 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_dequad_qg_trimmed_75000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_dequad_qg_trimmed_75000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_dequad_qg_trimmed_75000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_dequad_qg_trimmed_75000_en_5.4.2_3.0_1723943999051.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_dequad_qg_trimmed_75000_en_5.4.2_3.0_1723943999051.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_dequad_qg_trimmed_75000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_dequad_qg_trimmed_75000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_dequad_qg_trimmed_75000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-dequad-qg-trimmed-75000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_pipeline_en.md new file mode 100644 index 00000000000000..fd8b4b58b4febb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_dequad_qg_trimmed_75000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_dequad_qg_trimmed_75000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_dequad_qg_trimmed_75000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_dequad_qg_trimmed_75000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_dequad_qg_trimmed_75000_pipeline_en_5.4.2_3.0_1723944069002.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_dequad_qg_trimmed_75000_pipeline_en_5.4.2_3.0_1723944069002.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_dequad_qg_trimmed_75000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_dequad_qg_trimmed_75000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_dequad_qg_trimmed_75000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-dequad-qg-trimmed-75000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_esquad_qg_ae_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_esquad_qg_ae_pipeline_es.md new file mode 100644 index 00000000000000..b654bf93d1a8c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_esquad_qg_ae_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_base_esquad_qg_ae_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_esquad_qg_ae_pipeline +date: 2024-08-18 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_esquad_qg_ae_pipeline` is a Castilian, Spanish model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_esquad_qg_ae_pipeline_es_5.4.2_3.0_1724022405875.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_esquad_qg_ae_pipeline_es_5.4.2_3.0_1724022405875.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_esquad_qg_ae_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_esquad_qg_ae_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_esquad_qg_ae_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|2.4 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-esquad-qg-ae + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister_en.md new file mode 100644 index 00000000000000..662f90c1fcef25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister T5Transformer from himanshubeniwal +author: John Snow Labs +name: mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister` is a English model originally trained by himanshubeniwal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister_en_5.4.2_3.0_1723997857561.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister_en_5.4.2_3.0_1723997857561.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_gujarati_tonga_tonga_islands_english_filthy_primeminister| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/himanshubeniwal/mt5-base-finetuned-gu-to-en-filthy-PrimeMinister \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_en.md new file mode 100644 index 00000000000000..e3633bde87fc39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_mt5_base_poem_hemen T5Transformer from shkna1368 +author: John Snow Labs +name: mt5_base_finetuned_mt5_base_poem_hemen +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_mt5_base_poem_hemen` is a English model originally trained by shkna1368. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_mt5_base_poem_hemen_en_5.4.2_3.0_1723954403740.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_mt5_base_poem_hemen_en_5.4.2_3.0_1723954403740.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_mt5_base_poem_hemen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_mt5_base_poem_hemen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_mt5_base_poem_hemen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/shkna1368/mt5-base-finetuned-mt5-base-poem-hemen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_pipeline_en.md new file mode 100644 index 00000000000000..7203323793736e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_mt5_base_poem_hemen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_finetuned_mt5_base_poem_hemen_pipeline pipeline T5Transformer from shkna1368 +author: John Snow Labs +name: mt5_base_finetuned_mt5_base_poem_hemen_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_mt5_base_poem_hemen_pipeline` is a English model originally trained by shkna1368. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_mt5_base_poem_hemen_pipeline_en_5.4.2_3.0_1723954662745.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_mt5_base_poem_hemen_pipeline_en_5.4.2_3.0_1723954662745.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_finetuned_mt5_base_poem_hemen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_finetuned_mt5_base_poem_hemen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_mt5_base_poem_hemen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/shkna1368/mt5-base-finetuned-mt5-base-poem-hemen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base_en.md new file mode 100644 index 00000000000000..f86636b814fd14 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base T5Transformer from nestoralvaro +author: John Snow Labs +name: mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base` is a English model originally trained by nestoralvaro. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base_en_5.4.2_3.0_1723943701259.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base_en_5.4.2_3.0_1723943701259.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_xsum_raw_data_prep_2021_12_26___t22027_162754_csv__google_mt5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/nestoralvaro/mt5-base-finetuned-xsum-RAW_data_prep_2021_12_26___t22027_162754.csv__google_mt5_base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_en.md new file mode 100644 index 00000000000000..293caeb1b6d631 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_gecid_e8_b8 T5Transformer from jeremyvictor +author: John Snow Labs +name: mt5_base_gecid_e8_b8 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_gecid_e8_b8` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_gecid_e8_b8_en_5.4.2_3.0_1724004768126.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_gecid_e8_b8_en_5.4.2_3.0_1724004768126.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_gecid_e8_b8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_gecid_e8_b8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_gecid_e8_b8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/jeremyvictor/mt5-base-gecid-e8-b8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_pipeline_en.md new file mode 100644 index 00000000000000..52d0d4b98454ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_gecid_e8_b8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_gecid_e8_b8_pipeline pipeline T5Transformer from jeremyvictor +author: John Snow Labs +name: mt5_base_gecid_e8_b8_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_gecid_e8_b8_pipeline` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_gecid_e8_b8_pipeline_en_5.4.2_3.0_1724005070327.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_gecid_e8_b8_pipeline_en_5.4.2_3.0_1724005070327.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_gecid_e8_b8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_gecid_e8_b8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_gecid_e8_b8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/jeremyvictor/mt5-base-gecid-e8-b8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_en.md new file mode 100644 index 00000000000000..b441dd51a822cb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_intermediate_4_merged T5Transformer from Sunbird +author: John Snow Labs +name: mt5_base_intermediate_4_merged +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_intermediate_4_merged` is a English model originally trained by Sunbird. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_intermediate_4_merged_en_5.4.2_3.0_1723979341499.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_intermediate_4_merged_en_5.4.2_3.0_1723979341499.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_intermediate_4_merged","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_intermediate_4_merged", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_intermediate_4_merged| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/Sunbird/mt5-base-intermediate-4-merged \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_pipeline_en.md new file mode 100644 index 00000000000000..b80e3a54b3e12e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_intermediate_4_merged_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_intermediate_4_merged_pipeline pipeline T5Transformer from Sunbird +author: John Snow Labs +name: mt5_base_intermediate_4_merged_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_intermediate_4_merged_pipeline` is a English model originally trained by Sunbird. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_intermediate_4_merged_pipeline_en_5.4.2_3.0_1723979519266.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_intermediate_4_merged_pipeline_en_5.4.2_3.0_1723979519266.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_intermediate_4_merged_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_intermediate_4_merged_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_intermediate_4_merged_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/Sunbird/mt5-base-intermediate-4-merged + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_en.md new file mode 100644 index 00000000000000..b66f74d889a871 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_itquad_qg_ae_trimmed_50000 T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_itquad_qg_ae_trimmed_50000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_itquad_qg_ae_trimmed_50000` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_itquad_qg_ae_trimmed_50000_en_5.4.2_3.0_1723998468878.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_itquad_qg_ae_trimmed_50000_en_5.4.2_3.0_1723998468878.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_itquad_qg_ae_trimmed_50000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_itquad_qg_ae_trimmed_50000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_itquad_qg_ae_trimmed_50000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-itquad-qg-ae-trimmed-50000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_pipeline_en.md new file mode 100644 index 00000000000000..60aed31d03d23b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_itquad_qg_ae_trimmed_50000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_itquad_qg_ae_trimmed_50000_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_itquad_qg_ae_trimmed_50000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_itquad_qg_ae_trimmed_50000_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_itquad_qg_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1723998522883.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_itquad_qg_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1723998522883.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_itquad_qg_ae_trimmed_50000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_itquad_qg_ae_trimmed_50000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_itquad_qg_ae_trimmed_50000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-itquad-qg-ae-trimmed-50000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_en.md new file mode 100644 index 00000000000000..1d835a88cc9f37 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_jaquad_qg_trimmed_90000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_jaquad_qg_trimmed_90000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qg_trimmed_90000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_90000_en_5.4.2_3.0_1723997029492.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_90000_en_5.4.2_3.0_1723997029492.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_jaquad_qg_trimmed_90000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_jaquad_qg_trimmed_90000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qg_trimmed_90000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.4 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-jaquad-qg-trimmed-90000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_pipeline_en.md new file mode 100644 index 00000000000000..ba7ec797090ae9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_jaquad_qg_trimmed_90000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_jaquad_qg_trimmed_90000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_jaquad_qg_trimmed_90000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qg_trimmed_90000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_90000_pipeline_en_5.4.2_3.0_1723997110328.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_90000_pipeline_en_5.4.2_3.0_1723997110328.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_jaquad_qg_trimmed_90000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_jaquad_qg_trimmed_90000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qg_trimmed_90000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.4 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-jaquad-qg-trimmed-90000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_en.md new file mode 100644 index 00000000000000..50e9c712f3fcd6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_nc16_10k_deen T5Transformer from leukas +author: John Snow Labs +name: mt5_base_nc16_10k_deen +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_nc16_10k_deen` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_10k_deen_en_5.4.2_3.0_1723972835547.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_10k_deen_en_5.4.2_3.0_1723972835547.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_nc16_10k_deen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_nc16_10k_deen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_nc16_10k_deen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/leukas/mt5-base-nc16-10k-deen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_pipeline_en.md new file mode 100644 index 00000000000000..5c1c6efc476ded --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_10k_deen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_nc16_10k_deen_pipeline pipeline T5Transformer from leukas +author: John Snow Labs +name: mt5_base_nc16_10k_deen_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_nc16_10k_deen_pipeline` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_10k_deen_pipeline_en_5.4.2_3.0_1723973143088.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_10k_deen_pipeline_en_5.4.2_3.0_1723973143088.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_nc16_10k_deen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_nc16_10k_deen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_nc16_10k_deen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/leukas/mt5-base-nc16-10k-deen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_2k_ruen_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_2k_ruen_en.md new file mode 100644 index 00000000000000..2c431d54d569dd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_nc16_2k_ruen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_nc16_2k_ruen T5Transformer from leukas +author: John Snow Labs +name: mt5_base_nc16_2k_ruen +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_nc16_2k_ruen` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_2k_ruen_en_5.4.2_3.0_1723976108084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_nc16_2k_ruen_en_5.4.2_3.0_1723976108084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_nc16_2k_ruen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_nc16_2k_ruen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_nc16_2k_ruen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/leukas/mt5-base-nc16-2k-ruen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_en.md new file mode 100644 index 00000000000000..91175f6d0609ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_ruquad_qg_trimmed_60000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_ruquad_qg_trimmed_60000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ruquad_qg_trimmed_60000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ruquad_qg_trimmed_60000_en_5.4.2_3.0_1723987590043.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ruquad_qg_trimmed_60000_en_5.4.2_3.0_1723987590043.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_ruquad_qg_trimmed_60000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_ruquad_qg_trimmed_60000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ruquad_qg_trimmed_60000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-ruquad-qg-trimmed-60000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_pipeline_en.md new file mode 100644 index 00000000000000..f26cbca39f2e73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_ruquad_qg_trimmed_60000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_ruquad_qg_trimmed_60000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_ruquad_qg_trimmed_60000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ruquad_qg_trimmed_60000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ruquad_qg_trimmed_60000_pipeline_en_5.4.2_3.0_1723987655786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ruquad_qg_trimmed_60000_pipeline_en_5.4.2_3.0_1723987655786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_ruquad_qg_trimmed_60000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_ruquad_qg_trimmed_60000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ruquad_qg_trimmed_60000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-ruquad-qg-trimmed-60000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_en.md new file mode 100644 index 00000000000000..fd0e6aa479b683 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_trimmed_russian T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_russian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_russian` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_russian_en_5.4.2_3.0_1723972143790.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_russian_en_5.4.2_3.0_1723972143790.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_trimmed_russian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_trimmed_russian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_russian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-ru \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_pipeline_en.md new file mode 100644 index 00000000000000..1b0df36135ef97 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_russian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_trimmed_russian_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_russian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_russian_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_russian_pipeline_en_5.4.2_3.0_1723972491292.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_russian_pipeline_en_5.4.2_3.0_1723972491292.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_trimmed_russian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_trimmed_russian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_russian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-ru + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_en.md new file mode 100644 index 00000000000000..fc738ecc515123 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_trimmed_spanish_45000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_spanish_45000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_spanish_45000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_spanish_45000_en_5.4.2_3.0_1724004442377.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_spanish_45000_en_5.4.2_3.0_1724004442377.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_trimmed_spanish_45000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_trimmed_spanish_45000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_spanish_45000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|579.6 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-es-45000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_pipeline_en.md new file mode 100644 index 00000000000000..45ffb0a5525941 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_base_trimmed_spanish_45000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_trimmed_spanish_45000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_spanish_45000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_spanish_45000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_spanish_45000_pipeline_en_5.4.2_3.0_1724004634184.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_spanish_45000_pipeline_en_5.4.2_3.0_1724004634184.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_trimmed_spanish_45000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_trimmed_spanish_45000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_spanish_45000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|579.6 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-es-45000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_en.md new file mode 100644 index 00000000000000..ef7861501e8a52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_empty_desc_5k_msp T5Transformer from Roy029 +author: John Snow Labs +name: mt5_empty_desc_5k_msp +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_empty_desc_5k_msp` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_empty_desc_5k_msp_en_5.4.2_3.0_1723980774955.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_empty_desc_5k_msp_en_5.4.2_3.0_1723980774955.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_empty_desc_5k_msp","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_empty_desc_5k_msp", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_empty_desc_5k_msp| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_empty_desc_5k_msp \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_pipeline_en.md new file mode 100644 index 00000000000000..dbedddc6c2f70b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_empty_desc_5k_msp_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_empty_desc_5k_msp_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: mt5_empty_desc_5k_msp_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_empty_desc_5k_msp_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_empty_desc_5k_msp_pipeline_en_5.4.2_3.0_1723980962572.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_empty_desc_5k_msp_pipeline_en_5.4.2_3.0_1723980962572.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_empty_desc_5k_msp_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_empty_desc_5k_msp_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_empty_desc_5k_msp_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_empty_desc_5k_msp + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_english_rr_50_norwegian_bokml_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_rr_50_norwegian_bokml_en.md new file mode 100644 index 00000000000000..c048e98ceec533 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_rr_50_norwegian_bokml_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_english_rr_50_norwegian_bokml T5Transformer from MarianaLC +author: John Snow Labs +name: mt5_english_rr_50_norwegian_bokml +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_english_rr_50_norwegian_bokml` is a English model originally trained by MarianaLC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_english_rr_50_norwegian_bokml_en_5.4.2_3.0_1723960980088.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_english_rr_50_norwegian_bokml_en_5.4.2_3.0_1723960980088.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_english_rr_50_norwegian_bokml","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_english_rr_50_norwegian_bokml", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_english_rr_50_norwegian_bokml| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/MarianaLC/mt5-en-rr-50-nb \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_en.md new file mode 100644 index 00000000000000..1028511408c235 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_english_tonga_tonga_islands_hindi_v2 T5Transformer from jash33 +author: John Snow Labs +name: mt5_english_tonga_tonga_islands_hindi_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_english_tonga_tonga_islands_hindi_v2` is a English model originally trained by jash33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_english_tonga_tonga_islands_hindi_v2_en_5.4.2_3.0_1723995568457.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_english_tonga_tonga_islands_hindi_v2_en_5.4.2_3.0_1723995568457.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_english_tonga_tonga_islands_hindi_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_english_tonga_tonga_islands_hindi_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_english_tonga_tonga_islands_hindi_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/jash33/mt5-en-to-hi-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_pipeline_en.md new file mode 100644 index 00000000000000..90ce4b47420ec0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_english_tonga_tonga_islands_hindi_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_english_tonga_tonga_islands_hindi_v2_pipeline pipeline T5Transformer from jash33 +author: John Snow Labs +name: mt5_english_tonga_tonga_islands_hindi_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_english_tonga_tonga_islands_hindi_v2_pipeline` is a English model originally trained by jash33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_english_tonga_tonga_islands_hindi_v2_pipeline_en_5.4.2_3.0_1723995662865.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_english_tonga_tonga_islands_hindi_v2_pipeline_en_5.4.2_3.0_1723995662865.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_english_tonga_tonga_islands_hindi_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_english_tonga_tonga_islands_hindi_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_english_tonga_tonga_islands_hindi_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/jash33/mt5-en-to-hi-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_en.md new file mode 100644 index 00000000000000..09fc51c07abe9e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_extend_2500_nepal_bhasa T5Transformer from Roy029 +author: John Snow Labs +name: mt5_extend_2500_nepal_bhasa +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_extend_2500_nepal_bhasa` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_extend_2500_nepal_bhasa_en_5.4.2_3.0_1723942856605.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_extend_2500_nepal_bhasa_en_5.4.2_3.0_1723942856605.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_extend_2500_nepal_bhasa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_extend_2500_nepal_bhasa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_extend_2500_nepal_bhasa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|835.0 MB| + +## References + +https://huggingface.co/Roy029/mt5_extend_2500_new \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_pipeline_en.md new file mode 100644 index 00000000000000..f0301bab1b59ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_extend_2500_nepal_bhasa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_extend_2500_nepal_bhasa_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: mt5_extend_2500_nepal_bhasa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_extend_2500_nepal_bhasa_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_extend_2500_nepal_bhasa_pipeline_en_5.4.2_3.0_1723943144876.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_extend_2500_nepal_bhasa_pipeline_en_5.4.2_3.0_1723943144876.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_extend_2500_nepal_bhasa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_extend_2500_nepal_bhasa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_extend_2500_nepal_bhasa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|835.0 MB| + +## References + +https://huggingface.co/Roy029/mt5_extend_2500_new + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_korean_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_korean_en.md new file mode 100644 index 00000000000000..7491cac6222a78 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_korean_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_korean T5Transformer from bibekyess +author: John Snow Labs +name: mt5_korean +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_korean` is a English model originally trained by bibekyess. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_korean_en_5.4.2_3.0_1723942618405.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_korean_en_5.4.2_3.0_1723942618405.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_korean","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_korean", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_korean| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/bibekyess/mt5-korean \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_en.md new file mode 100644 index 00000000000000..ad5de1b252909f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_orig_human T5Transformer from Roy029 +author: John Snow Labs +name: mt5_orig_human +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_orig_human` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_orig_human_en_5.4.2_3.0_1724000293353.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_orig_human_en_5.4.2_3.0_1724000293353.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_orig_human","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_orig_human", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_orig_human| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_orig_human \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_pipeline_en.md new file mode 100644 index 00000000000000..bff7f3f787b5c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_orig_human_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_orig_human_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: mt5_orig_human_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_orig_human_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_orig_human_pipeline_en_5.4.2_3.0_1724000471515.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_orig_human_pipeline_en_5.4.2_3.0_1724000471515.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_orig_human_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_orig_human_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_orig_human_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_orig_human + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_en.md new file mode 100644 index 00000000000000..51da45a234ce27 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_amazon_title_generation_japanese T5Transformer from rkamimae +author: John Snow Labs +name: mt5_small_amazon_title_generation_japanese +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_amazon_title_generation_japanese` is a English model originally trained by rkamimae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_amazon_title_generation_japanese_en_5.4.2_3.0_1723961275767.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_amazon_title_generation_japanese_en_5.4.2_3.0_1723961275767.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_amazon_title_generation_japanese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_amazon_title_generation_japanese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_amazon_title_generation_japanese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/rkamimae/mt5-small-amazon-title-generation-japanese \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_pipeline_en.md new file mode 100644 index 00000000000000..31257f3ba57cec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_amazon_title_generation_japanese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_amazon_title_generation_japanese_pipeline pipeline T5Transformer from rkamimae +author: John Snow Labs +name: mt5_small_amazon_title_generation_japanese_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_amazon_title_generation_japanese_pipeline` is a English model originally trained by rkamimae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_amazon_title_generation_japanese_pipeline_en_5.4.2_3.0_1723961369948.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_amazon_title_generation_japanese_pipeline_en_5.4.2_3.0_1723961369948.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_amazon_title_generation_japanese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_amazon_title_generation_japanese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_amazon_title_generation_japanese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/rkamimae/mt5-small-amazon-title-generation-japanese + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_en.md new file mode 100644 index 00000000000000..6db5f53e32259a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_azerbaijani_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_azerbaijani_10k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_azerbaijani_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_azerbaijani_10k_en_5.4.2_3.0_1723961065886.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_azerbaijani_10k_en_5.4.2_3.0_1723961065886.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_azerbaijani_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_azerbaijani_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_azerbaijani_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-az-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_pipeline_en.md new file mode 100644 index 00000000000000..9fd1629f3198c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_azerbaijani_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_azerbaijani_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_azerbaijani_10k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_azerbaijani_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_azerbaijani_10k_pipeline_en_5.4.2_3.0_1723961244128.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_azerbaijani_10k_pipeline_en_5.4.2_3.0_1723961244128.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_azerbaijani_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_azerbaijani_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_azerbaijani_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-az-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_en.md new file mode 100644 index 00000000000000..6383393523b7ff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_belarusian_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_belarusian_10k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_belarusian_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_belarusian_10k_en_5.4.2_3.0_1724003717253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_belarusian_10k_en_5.4.2_3.0_1724003717253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_belarusian_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_belarusian_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_belarusian_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-be-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_pipeline_en.md new file mode 100644 index 00000000000000..88796e27176b2f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_belarusian_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_belarusian_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_belarusian_10k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_belarusian_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_belarusian_10k_pipeline_en_5.4.2_3.0_1724003859877.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_belarusian_10k_pipeline_en_5.4.2_3.0_1724003859877.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_belarusian_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_belarusian_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_belarusian_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-be-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_en.md new file mode 100644 index 00000000000000..13ae096ed8a30f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_codesearchnet_python3 T5Transformer from lmeninato +author: John Snow Labs +name: mt5_small_codesearchnet_python3 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_codesearchnet_python3` is a English model originally trained by lmeninato. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_codesearchnet_python3_en_5.4.2_3.0_1723982964164.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_codesearchnet_python3_en_5.4.2_3.0_1723982964164.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_codesearchnet_python3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_codesearchnet_python3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_codesearchnet_python3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/lmeninato/mt5-small-codesearchnet-python3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_pipeline_en.md new file mode 100644 index 00000000000000..c092a220270f43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_codesearchnet_python3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_codesearchnet_python3_pipeline pipeline T5Transformer from lmeninato +author: John Snow Labs +name: mt5_small_codesearchnet_python3_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_codesearchnet_python3_pipeline` is a English model originally trained by lmeninato. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_codesearchnet_python3_pipeline_en_5.4.2_3.0_1723983250448.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_codesearchnet_python3_pipeline_en_5.4.2_3.0_1723983250448.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_codesearchnet_python3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_codesearchnet_python3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_codesearchnet_python3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/lmeninato/mt5-small-codesearchnet-python3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_es.md new file mode 100644 index 00000000000000..24bf3d0a93e682 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_esquad_qa T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_esquad_qa +date: 2024-08-18 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_esquad_qa` is a Castilian, Spanish model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_esquad_qa_es_5.4.2_3.0_1723987206039.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_esquad_qa_es_5.4.2_3.0_1723987206039.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_esquad_qa","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_esquad_qa", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_esquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|1.3 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-esquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_pipeline_es.md new file mode 100644 index 00000000000000..f0e0a234eae6ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_esquad_qa_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_esquad_qa_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_esquad_qa_pipeline +date: 2024-08-18 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_esquad_qa_pipeline` is a Castilian, Spanish model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_esquad_qa_pipeline_es_5.4.2_3.0_1723987292512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_esquad_qa_pipeline_es_5.4.2_3.0_1723987292512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_esquad_qa_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_esquad_qa_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_esquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|1.3 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-esquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_en.md new file mode 100644 index 00000000000000..0561f997c2ef7d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian T5Transformer from mriggs +author: John Snow Labs +name: mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian` is a English model originally trained by mriggs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724001356548.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724001356548.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/mriggs/mt5-small-finetuned-8epochs-opus_books-en-to-it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline_en.md new file mode 100644 index 00000000000000..dc615f6285b2ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline pipeline T5Transformer from mriggs +author: John Snow Labs +name: mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline` is a English model originally trained by mriggs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724001444465.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724001444465.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_8epochs_opus_books_english_tonga_tonga_islands_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/mriggs/mt5-small-finetuned-8epochs-opus_books-en-to-it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_en.md new file mode 100644 index 00000000000000..8e1280fb010c9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_alexandermoshnov T5Transformer from AlexanderMoshnov +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_alexandermoshnov +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_alexandermoshnov` is a English model originally trained by AlexanderMoshnov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_alexandermoshnov_en_5.4.2_3.0_1724006631793.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_alexandermoshnov_en_5.4.2_3.0_1724006631793.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_alexandermoshnov","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_alexandermoshnov", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_alexandermoshnov| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AlexanderMoshnov/mt5-small-finetuned-amazon-en-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline_en.md new file mode 100644 index 00000000000000..a5c15b5c207b10 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline pipeline T5Transformer from AlexanderMoshnov +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline` is a English model originally trained by AlexanderMoshnov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline_en_5.4.2_3.0_1724006751123.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline_en_5.4.2_3.0_1724006751123.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_alexandermoshnov_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AlexanderMoshnov/mt5-small-finetuned-amazon-en-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_en.md new file mode 100644 index 00000000000000..2bb7c8cc4d56cc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_gsvann T5Transformer from gsvann +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_gsvann +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_gsvann` is a English model originally trained by gsvann. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_gsvann_en_5.4.2_3.0_1724002452002.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_gsvann_en_5.4.2_3.0_1724002452002.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_gsvann","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_gsvann", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_gsvann| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/gsvann/mt5-small-finetuned-amazon-en-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_pipeline_en.md new file mode 100644 index 00000000000000..15af3891a54e41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_german_gsvann_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_gsvann_pipeline pipeline T5Transformer from gsvann +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_gsvann_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_gsvann_pipeline` is a English model originally trained by gsvann. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_gsvann_pipeline_en_5.4.2_3.0_1724002540625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_gsvann_pipeline_en_5.4.2_3.0_1724002540625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_german_gsvann_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_german_gsvann_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_gsvann_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/gsvann/mt5-small-finetuned-amazon-en-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_en.md new file mode 100644 index 00000000000000..c2866ed0811ad9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_juanarturovargas T5Transformer from juanarturovargas +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_juanarturovargas +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_juanarturovargas` is a English model originally trained by juanarturovargas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_juanarturovargas_en_5.4.2_3.0_1723977786489.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_juanarturovargas_en_5.4.2_3.0_1723977786489.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_juanarturovargas","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_juanarturovargas", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_juanarturovargas| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/juanarturovargas/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline_en.md new file mode 100644 index 00000000000000..459b4d59e6dfcc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline pipeline T5Transformer from juanarturovargas +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline` is a English model originally trained by juanarturovargas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline_en_5.4.2_3.0_1723977891484.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline_en_5.4.2_3.0_1723977891484.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_juanarturovargas_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/juanarturovargas/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_en.md new file mode 100644 index 00000000000000..1a21c77656dd99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_wujia T5Transformer from wujia +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_wujia +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_wujia` is a English model originally trained by wujia. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_wujia_en_5.4.2_3.0_1723966006026.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_wujia_en_5.4.2_3.0_1723966006026.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_wujia","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_wujia", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_wujia| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/wujia/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_pipeline_en.md new file mode 100644 index 00000000000000..1185ad5a3d8f68 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_amazon_english_spanish_wujia_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_wujia_pipeline pipeline T5Transformer from wujia +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_wujia_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_wujia_pipeline` is a English model originally trained by wujia. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_wujia_pipeline_en_5.4.2_3.0_1723966122839.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_wujia_pipeline_en_5.4.2_3.0_1723966122839.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_wujia_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_wujia_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_wujia_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/wujia/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_en.md new file mode 100644 index 00000000000000..763b210c40aa0e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_english_tonga_tonga_islands_romanian T5Transformer from laymanyet +author: John Snow Labs +name: mt5_small_finetuned_english_tonga_tonga_islands_romanian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_english_tonga_tonga_islands_romanian` is a English model originally trained by laymanyet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_romanian_en_5.4.2_3.0_1723950551756.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_romanian_en_5.4.2_3.0_1723950551756.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_english_tonga_tonga_islands_romanian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_english_tonga_tonga_islands_romanian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_english_tonga_tonga_islands_romanian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/laymanyet/mt5-small-finetuned-en-to-ro \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline_en.md new file mode 100644 index 00000000000000..9ed540e7cee935 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline pipeline T5Transformer from laymanyet +author: John Snow Labs +name: mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline` is a English model originally trained by laymanyet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline_en_5.4.2_3.0_1723950841786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline_en_5.4.2_3.0_1723950841786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_english_tonga_tonga_islands_romanian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/laymanyet/mt5-small-finetuned-en-to-ro + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_en.md new file mode 100644 index 00000000000000..874569d8436d91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_test_fqw T5Transformer from fqw +author: John Snow Labs +name: mt5_small_finetuned_test_fqw +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_test_fqw` is a English model originally trained by fqw. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_test_fqw_en_5.4.2_3.0_1723965350279.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_test_fqw_en_5.4.2_3.0_1723965350279.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_test_fqw","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_test_fqw", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_test_fqw| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/fqw/mt5-small-finetuned-test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_pipeline_en.md new file mode 100644 index 00000000000000..d7132dfec6ccf3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_finetuned_test_fqw_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_test_fqw_pipeline pipeline T5Transformer from fqw +author: John Snow Labs +name: mt5_small_finetuned_test_fqw_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_test_fqw_pipeline` is a English model originally trained by fqw. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_test_fqw_pipeline_en_5.4.2_3.0_1723965635324.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_test_fqw_pipeline_en_5.4.2_3.0_1723965635324.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_test_fqw_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_test_fqw_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_test_fqw_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/fqw/mt5-small-finetuned-test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_en.md new file mode 100644 index 00000000000000..40c721b3956d4f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_french_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_french_10k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_french_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_french_10k_en_5.4.2_3.0_1723954571600.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_french_10k_en_5.4.2_3.0_1723954571600.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_french_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_french_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_french_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-fr-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_pipeline_en.md new file mode 100644 index 00000000000000..1af05d4348d796 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_french_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_french_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_french_10k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_french_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_french_10k_pipeline_en_5.4.2_3.0_1723954723613.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_french_10k_pipeline_en_5.4.2_3.0_1723954723613.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_french_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_french_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_french_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-fr-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_en.md new file mode 100644 index 00000000000000..54e52f681c31e7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_jaquad_qa_trimmed_japanese T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qa_trimmed_japanese +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qa_trimmed_japanese` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_en_5.4.2_3.0_1724010610859.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_en_5.4.2_3.0_1724010610859.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_jaquad_qa_trimmed_japanese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_jaquad_qa_trimmed_japanese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qa_trimmed_japanese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|728.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qa-trimmed-ja \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_pipeline_en.md new file mode 100644 index 00000000000000..9a6a3d404e89e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qa_trimmed_japanese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_jaquad_qa_trimmed_japanese_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qa_trimmed_japanese_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qa_trimmed_japanese_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_pipeline_en_5.4.2_3.0_1724010656838.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_pipeline_en_5.4.2_3.0_1724010656838.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_jaquad_qa_trimmed_japanese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_jaquad_qa_trimmed_japanese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qa_trimmed_japanese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|728.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qa-trimmed-ja + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_en.md new file mode 100644 index 00000000000000..84844e360513aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_jaquad_qg_trimmed_japanese_5000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qg_trimmed_japanese_5000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qg_trimmed_japanese_5000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qg_trimmed_japanese_5000_en_5.4.2_3.0_1723962235004.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qg_trimmed_japanese_5000_en_5.4.2_3.0_1723962235004.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_jaquad_qg_trimmed_japanese_5000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_jaquad_qg_trimmed_japanese_5000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qg_trimmed_japanese_5000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|195.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qg-trimmed-ja-5000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline_en.md new file mode 100644 index 00000000000000..2e114d3cdf2227 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline_en_5.4.2_3.0_1723962244549.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline_en_5.4.2_3.0_1723962244549.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qg_trimmed_japanese_5000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|195.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qg-trimmed-ja-5000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_ko.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_ko.md new file mode 100644 index 00000000000000..f00e4d0fd0302f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_ko.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Korean mt5_small_koquad_qag T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_koquad_qag +date: 2024-08-18 +tags: [ko, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_koquad_qag` is a Korean model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qag_ko_5.4.2_3.0_1724019542360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qag_ko_5.4.2_3.0_1724019542360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_koquad_qag","ko") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_koquad_qag", "ko") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_koquad_qag| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ko| +|Size:|1.2 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-koquad-qag \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_pipeline_ko.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_pipeline_ko.md new file mode 100644 index 00000000000000..642168d4f25529 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_koquad_qag_pipeline_ko.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Korean mt5_small_koquad_qag_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_koquad_qag_pipeline +date: 2024-08-18 +tags: [ko, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_koquad_qag_pipeline` is a Korean model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qag_pipeline_ko_5.4.2_3.0_1724019634757.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qag_pipeline_ko_5.4.2_3.0_1724019634757.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_koquad_qag_pipeline", lang = "ko") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_koquad_qag_pipeline", lang = "ko") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_koquad_qag_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ko| +|Size:|1.2 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-koquad-qag + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_en.md new file mode 100644 index 00000000000000..0e426b4bde6414 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_nc16_10k_ptes T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_10k_ptes +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_10k_ptes` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_10k_ptes_en_5.4.2_3.0_1723949815692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_10k_ptes_en_5.4.2_3.0_1723949815692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_nc16_10k_ptes","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_nc16_10k_ptes", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_10k_ptes| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-10k-ptes \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_pipeline_en.md new file mode 100644 index 00000000000000..46c775e9556ca8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_nc16_10k_ptes_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_nc16_10k_ptes_pipeline pipeline T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_10k_ptes_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_10k_ptes_pipeline` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_10k_ptes_pipeline_en_5.4.2_3.0_1723950016051.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_10k_ptes_pipeline_en_5.4.2_3.0_1723950016051.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_nc16_10k_ptes_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_nc16_10k_ptes_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_10k_ptes_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-10k-ptes + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_en.md new file mode 100644 index 00000000000000..d3ec526c236fe1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_normail T5Transformer from bg79-v23-bidata-ntnu +author: John Snow Labs +name: mt5_small_normail +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_normail` is a English model originally trained by bg79-v23-bidata-ntnu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_normail_en_5.4.2_3.0_1723957050543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_normail_en_5.4.2_3.0_1723957050543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_normail","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_normail", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_normail| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/bg79-v23-bidata-ntnu/mt5_small-normail \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_pipeline_en.md new file mode 100644 index 00000000000000..35c163eab2f339 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_normail_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_normail_pipeline pipeline T5Transformer from bg79-v23-bidata-ntnu +author: John Snow Labs +name: mt5_small_normail_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_normail_pipeline` is a English model originally trained by bg79-v23-bidata-ntnu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_normail_pipeline_en_5.4.2_3.0_1723957150663.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_normail_pipeline_en_5.4.2_3.0_1723957150663.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_normail_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_normail_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_normail_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/bg79-v23-bidata-ntnu/mt5_small-normail + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_pipeline_th.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_pipeline_th.md new file mode 100644 index 00000000000000..a190f450dffc6a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_pipeline_th.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Thai mt5_small_simple_thai_keyword_2_text_generator_pipeline pipeline T5Transformer from c-tawayip +author: John Snow Labs +name: mt5_small_simple_thai_keyword_2_text_generator_pipeline +date: 2024-08-18 +tags: [th, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: th +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_simple_thai_keyword_2_text_generator_pipeline` is a Thai model originally trained by c-tawayip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_simple_thai_keyword_2_text_generator_pipeline_th_5.4.2_3.0_1723991790400.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_simple_thai_keyword_2_text_generator_pipeline_th_5.4.2_3.0_1723991790400.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_simple_thai_keyword_2_text_generator_pipeline", lang = "th") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_simple_thai_keyword_2_text_generator_pipeline", lang = "th") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_simple_thai_keyword_2_text_generator_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|th| +|Size:|1.2 GB| + +## References + +https://huggingface.co/c-tawayip/mt5-small-Simple-Thai-Keyword-2-Text-Generator + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_th.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_th.md new file mode 100644 index 00000000000000..3a07a51c0ac1bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_simple_thai_keyword_2_text_generator_th.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Thai mt5_small_simple_thai_keyword_2_text_generator T5Transformer from c-tawayip +author: John Snow Labs +name: mt5_small_simple_thai_keyword_2_text_generator +date: 2024-08-18 +tags: [th, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: th +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_simple_thai_keyword_2_text_generator` is a Thai model originally trained by c-tawayip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_simple_thai_keyword_2_text_generator_th_5.4.2_3.0_1723991694319.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_simple_thai_keyword_2_text_generator_th_5.4.2_3.0_1723991694319.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_simple_thai_keyword_2_text_generator","th") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_simple_thai_keyword_2_text_generator", "th") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_simple_thai_keyword_2_text_generator| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|th| +|Size:|1.2 GB| + +## References + +https://huggingface.co/c-tawayip/mt5-small-Simple-Thai-Keyword-2-Text-Generator \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_en.md new file mode 100644 index 00000000000000..e1df4b3ac0546d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_squad_qa_trimmed_english_30000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_squad_qa_trimmed_english_30000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qa_trimmed_english_30000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_30000_en_5.4.2_3.0_1724001845356.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_30000_en_5.4.2_3.0_1724001845356.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_squad_qa_trimmed_english_30000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_squad_qa_trimmed_english_30000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qa_trimmed_english_30000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-squad-qa-trimmed-en-30000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_pipeline_en.md new file mode 100644 index 00000000000000..8be6b7e816f733 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_squad_qa_trimmed_english_30000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_squad_qa_trimmed_english_30000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_squad_qa_trimmed_english_30000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qa_trimmed_english_30000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_30000_pipeline_en_5.4.2_3.0_1724001861182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_30000_pipeline_en_5.4.2_3.0_1724001861182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_squad_qa_trimmed_english_30000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_squad_qa_trimmed_english_30000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qa_trimmed_english_30000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-squad-qa-trimmed-en-30000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_en.md new file mode 100644 index 00000000000000..7980f4a5de5cb7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_task1_dataset3 T5Transformer from ZhiguangHan +author: John Snow Labs +name: mt5_small_task1_dataset3 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_task1_dataset3` is a English model originally trained by ZhiguangHan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_task1_dataset3_en_5.4.2_3.0_1723950812344.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_task1_dataset3_en_5.4.2_3.0_1723950812344.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_task1_dataset3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_task1_dataset3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_task1_dataset3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ZhiguangHan/mt5-small-task1-dataset3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_pipeline_en.md new file mode 100644 index 00000000000000..bc4ba45e9a3b46 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_task1_dataset3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_task1_dataset3_pipeline pipeline T5Transformer from ZhiguangHan +author: John Snow Labs +name: mt5_small_task1_dataset3_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_task1_dataset3_pipeline` is a English model originally trained by ZhiguangHan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_task1_dataset3_pipeline_en_5.4.2_3.0_1723950935923.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_task1_dataset3_pipeline_en_5.4.2_3.0_1723950935923.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_task1_dataset3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_task1_dataset3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_task1_dataset3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ZhiguangHan/mt5-small-task1-dataset3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_en.md new file mode 100644 index 00000000000000..eea36ca2495cbf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_thaisum_monthida T5Transformer from Monthida +author: John Snow Labs +name: mt5_small_thaisum_monthida +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_thaisum_monthida` is a English model originally trained by Monthida. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_thaisum_monthida_en_5.4.2_3.0_1723989359604.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_thaisum_monthida_en_5.4.2_3.0_1723989359604.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_thaisum_monthida","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_thaisum_monthida", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_thaisum_monthida| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Monthida/mt5-small-thaisum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_pipeline_en.md new file mode 100644 index 00000000000000..4a026286e28f18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_thaisum_monthida_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_thaisum_monthida_pipeline pipeline T5Transformer from Monthida +author: John Snow Labs +name: mt5_small_thaisum_monthida_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_thaisum_monthida_pipeline` is a English model originally trained by Monthida. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_thaisum_monthida_pipeline_en_5.4.2_3.0_1723989451151.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_thaisum_monthida_pipeline_en_5.4.2_3.0_1723989451151.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_thaisum_monthida_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_thaisum_monthida_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_thaisum_monthida_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Monthida/mt5-small-thaisum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_en.md new file mode 100644 index 00000000000000..3e83981d92f3d8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_english_60000_squad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_60000_squad_qa +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_60000_squad_qa` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qa_en_5.4.2_3.0_1723971186811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qa_en_5.4.2_3.0_1723971186811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_english_60000_squad_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_english_60000_squad_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_60000_squad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|485.0 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-60000-squad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_pipeline_en.md new file mode 100644 index 00000000000000..acb15f2578ee26 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_english_60000_squad_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_english_60000_squad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_60000_squad_qa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_60000_squad_qa_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qa_pipeline_en_5.4.2_3.0_1723971211472.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_60000_squad_qa_pipeline_en_5.4.2_3.0_1723971211472.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_english_60000_squad_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_english_60000_squad_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_60000_squad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|485.0 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-60000-squad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_en.md new file mode 100644 index 00000000000000..74c222a62906dd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_japanese_15000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_japanese_15000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_japanese_15000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_15000_en_5.4.2_3.0_1723991390152.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_15000_en_5.4.2_3.0_1723991390152.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_japanese_15000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_japanese_15000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_japanese_15000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|130.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ja-15000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_pipeline_en.md new file mode 100644 index 00000000000000..a5538a91fab9f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_japanese_15000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_japanese_15000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_japanese_15000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_japanese_15000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_15000_pipeline_en_5.4.2_3.0_1723991433227.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_15000_pipeline_en_5.4.2_3.0_1723991433227.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_japanese_15000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_japanese_15000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_japanese_15000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|130.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ja-15000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_pipeline_ru.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_pipeline_ru.md new file mode 100644 index 00000000000000..bda457180cc946 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_pipeline_ru.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_30000_ruquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_30000_ruquad_qg_pipeline +date: 2024-08-18 +tags: [ru, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_30000_ruquad_qg_pipeline` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_30000_ruquad_qg_pipeline_ru_5.4.2_3.0_1723991596103.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_30000_ruquad_qg_pipeline_ru_5.4.2_3.0_1723991596103.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_russian_30000_ruquad_qg_pipeline", lang = "ru") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_russian_30000_ruquad_qg_pipeline", lang = "ru") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_30000_ruquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ru| +|Size:|331.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-30000-ruquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_ru.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_ru.md new file mode 100644 index 00000000000000..69e6239b6a6edb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_30000_ruquad_qg_ru.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_30000_ruquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_30000_ruquad_qg +date: 2024-08-18 +tags: [ru, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_30000_ruquad_qg` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_30000_ruquad_qg_ru_5.4.2_3.0_1723991580360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_30000_ruquad_qg_ru_5.4.2_3.0_1723991580360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_30000_ruquad_qg","ru") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_30000_ruquad_qg", "ru") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_30000_ruquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ru| +|Size:|331.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-30000-ruquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_pipeline_ru.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_pipeline_ru.md new file mode 100644 index 00000000000000..c54255a335fede --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_pipeline_ru.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_90000_ruquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_90000_ruquad_qa_pipeline +date: 2024-08-18 +tags: [ru, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_90000_ruquad_qa_pipeline` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_90000_ruquad_qa_pipeline_ru_5.4.2_3.0_1723974495594.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_90000_ruquad_qa_pipeline_ru_5.4.2_3.0_1723974495594.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_russian_90000_ruquad_qa_pipeline", lang = "ru") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_russian_90000_ruquad_qa_pipeline", lang = "ru") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_90000_ruquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ru| +|Size:|592.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-90000-ruquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_ru.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_ru.md new file mode 100644 index 00000000000000..5d7670ce2e3586 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_russian_90000_ruquad_qa_ru.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_90000_ruquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_90000_ruquad_qa +date: 2024-08-18 +tags: [ru, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_90000_ruquad_qa` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_90000_ruquad_qa_ru_5.4.2_3.0_1723974460519.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_90000_ruquad_qa_ru_5.4.2_3.0_1723974460519.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_90000_ruquad_qa","ru") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_90000_ruquad_qa", "ru") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_90000_ruquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ru| +|Size:|592.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-90000-ruquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_es.md new file mode 100644 index 00000000000000..b45dc920e8b29a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_120000_esquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_120000_esquad_qa +date: 2024-08-18 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_120000_esquad_qa` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_120000_esquad_qa_es_5.4.2_3.0_1723946495498.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_120000_esquad_qa_es_5.4.2_3.0_1723946495498.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_120000_esquad_qa","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_120000_esquad_qa", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_120000_esquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|740.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-120000-esquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_pipeline_es.md new file mode 100644 index 00000000000000..53f01be90ce2a8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_120000_esquad_qa_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_120000_esquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_120000_esquad_qa_pipeline +date: 2024-08-18 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_120000_esquad_qa_pipeline` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_120000_esquad_qa_pipeline_es_5.4.2_3.0_1723946540481.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_120000_esquad_qa_pipeline_es_5.4.2_3.0_1723946540481.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_spanish_120000_esquad_qa_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_spanish_120000_esquad_qa_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_120000_esquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|740.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-120000-esquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_en.md new file mode 100644 index 00000000000000..6e39abd889c7cc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_spanish_5000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_5000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_5000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_5000_en_5.4.2_3.0_1723966575912.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_5000_en_5.4.2_3.0_1723966575912.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_5000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_5000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_5000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|101.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-5000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_pipeline_en.md new file mode 100644 index 00000000000000..dc8557495d8fc8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_5000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_spanish_5000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_5000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_5000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_5000_pipeline_en_5.4.2_3.0_1723966610724.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_5000_pipeline_en_5.4.2_3.0_1723966610724.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_spanish_5000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_spanish_5000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_5000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|101.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-5000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_es.md new file mode 100644 index 00000000000000..2cb010c273c9ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_60000_esquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_60000_esquad_qg +date: 2024-08-18 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_60000_esquad_qg` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_60000_esquad_qg_es_5.4.2_3.0_1724002827303.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_60000_esquad_qg_es_5.4.2_3.0_1724002827303.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_60000_esquad_qg","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_60000_esquad_qg", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_60000_esquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|476.0 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-60000-esquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_pipeline_es.md new file mode 100644 index 00000000000000..71a072514c06bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_small_trimmed_spanish_60000_esquad_qg_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_60000_esquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_60000_esquad_qg_pipeline +date: 2024-08-18 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_60000_esquad_qg_pipeline` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_60000_esquad_qg_pipeline_es_5.4.2_3.0_1724002853976.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_60000_esquad_qg_pipeline_es_5.4.2_3.0_1724002853976.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_spanish_60000_esquad_qg_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_spanish_60000_esquad_qg_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_60000_esquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|476.0 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-60000-esquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_en.md new file mode 100644 index 00000000000000..408e14f9e9c662 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_thaisum_finetune T5Transformer from Mayypeeya +author: John Snow Labs +name: mt5_thaisum_finetune +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thaisum_finetune` is a English model originally trained by Mayypeeya. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thaisum_finetune_en_5.4.2_3.0_1723953873331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thaisum_finetune_en_5.4.2_3.0_1723953873331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_thaisum_finetune","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_thaisum_finetune", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thaisum_finetune| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Mayypeeya/mt5_thaisum_finetune \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_pipeline_en.md new file mode 100644 index 00000000000000..5bb04d9083d5b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5_thaisum_finetune_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_thaisum_finetune_pipeline pipeline T5Transformer from Mayypeeya +author: John Snow Labs +name: mt5_thaisum_finetune_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thaisum_finetune_pipeline` is a English model originally trained by Mayypeeya. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thaisum_finetune_pipeline_en_5.4.2_3.0_1723954031807.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thaisum_finetune_pipeline_en_5.4.2_3.0_1723954031807.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_thaisum_finetune_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_thaisum_finetune_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thaisum_finetune_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Mayypeeya/mt5_thaisum_finetune + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_en.md new file mode 100644 index 00000000000000..caa8598e22d703 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5meu700 T5Transformer from mateiaassAI +author: John Snow Labs +name: mt5meu700 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5meu700` is a English model originally trained by mateiaassAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5meu700_en_5.4.2_3.0_1723957492062.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5meu700_en_5.4.2_3.0_1723957492062.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5meu700","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5meu700", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5meu700| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/mateiaassAI/mt5meu700 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_pipeline_en.md new file mode 100644 index 00000000000000..ee787c0fe93bb0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-mt5meu700_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5meu700_pipeline pipeline T5Transformer from mateiaassAI +author: John Snow Labs +name: mt5meu700_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5meu700_pipeline` is a English model originally trained by mateiaassAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5meu700_pipeline_en_5.4.2_3.0_1723957635689.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5meu700_pipeline_en_5.4.2_3.0_1723957635689.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5meu700_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5meu700_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5meu700_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/mateiaassAI/mt5meu700 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_en.md b/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_en.md new file mode 100644 index 00000000000000..cc71b5648d9a85 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English multi_news_train_4000t_t5_small_summary_model T5Transformer from AlexandraSnelling +author: John Snow Labs +name: multi_news_train_4000t_t5_small_summary_model +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`multi_news_train_4000t_t5_small_summary_model` is a English model originally trained by AlexandraSnelling. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/multi_news_train_4000t_t5_small_summary_model_en_5.4.2_3.0_1723964049792.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/multi_news_train_4000t_t5_small_summary_model_en_5.4.2_3.0_1723964049792.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("multi_news_train_4000t_t5_small_summary_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("multi_news_train_4000t_t5_small_summary_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|multi_news_train_4000t_t5_small_summary_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.1 MB| + +## References + +https://huggingface.co/AlexandraSnelling/multi_news_train_4000t_t5-small_summary_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_pipeline_en.md new file mode 100644 index 00000000000000..3a08b9b63dc89b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-multi_news_train_4000t_t5_small_summary_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English multi_news_train_4000t_t5_small_summary_model_pipeline pipeline T5Transformer from AlexandraSnelling +author: John Snow Labs +name: multi_news_train_4000t_t5_small_summary_model_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`multi_news_train_4000t_t5_small_summary_model_pipeline` is a English model originally trained by AlexandraSnelling. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/multi_news_train_4000t_t5_small_summary_model_pipeline_en_5.4.2_3.0_1723964069122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/multi_news_train_4000t_t5_small_summary_model_pipeline_en_5.4.2_3.0_1723964069122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("multi_news_train_4000t_t5_small_summary_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("multi_news_train_4000t_t5_small_summary_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|multi_news_train_4000t_t5_small_summary_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.1 MB| + +## References + +https://huggingface.co/AlexandraSnelling/multi_news_train_4000t_t5-small_summary_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_en.md b/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_en.md new file mode 100644 index 00000000000000..e1abc0bfdfffe2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nanot5_base T5Transformer from pnawrot +author: John Snow Labs +name: nanot5_base +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nanot5_base` is a English model originally trained by pnawrot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nanot5_base_en_5.4.2_3.0_1724022624713.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nanot5_base_en_5.4.2_3.0_1724022624713.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nanot5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nanot5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nanot5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pnawrot/nanoT5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_pipeline_en.md new file mode 100644 index 00000000000000..910bad6b84affb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nanot5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nanot5_base_pipeline pipeline T5Transformer from pnawrot +author: John Snow Labs +name: nanot5_base_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nanot5_base_pipeline` is a English model originally trained by pnawrot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nanot5_base_pipeline_en_5.4.2_3.0_1724022670895.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nanot5_base_pipeline_en_5.4.2_3.0_1724022670895.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nanot5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nanot5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nanot5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pnawrot/nanoT5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_en.md b/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_en.md new file mode 100644 index 00000000000000..ba2b4b889c59cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nativesql_schema T5Transformer from maaz77 +author: John Snow Labs +name: nativesql_schema +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nativesql_schema` is a English model originally trained by maaz77. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nativesql_schema_en_5.4.2_3.0_1723962752748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nativesql_schema_en_5.4.2_3.0_1723962752748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nativesql_schema","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nativesql_schema", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nativesql_schema| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/maaz77/NativeSQL_Schema \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_pipeline_en.md new file mode 100644 index 00000000000000..b3929f6bf5fb59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nativesql_schema_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nativesql_schema_pipeline pipeline T5Transformer from maaz77 +author: John Snow Labs +name: nativesql_schema_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nativesql_schema_pipeline` is a English model originally trained by maaz77. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nativesql_schema_pipeline_en_5.4.2_3.0_1723962800261.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nativesql_schema_pipeline_en_5.4.2_3.0_1723962800261.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nativesql_schema_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nativesql_schema_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nativesql_schema_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/maaz77/NativeSQL_Schema + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_en.md new file mode 100644 index 00000000000000..8aa8b6a71bf715 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ndc_summarization_v1 T5Transformer from huyenquinn282 +author: John Snow Labs +name: ndc_summarization_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ndc_summarization_v1` is a English model originally trained by huyenquinn282. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ndc_summarization_v1_en_5.4.2_3.0_1723969409512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ndc_summarization_v1_en_5.4.2_3.0_1723969409512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ndc_summarization_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ndc_summarization_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ndc_summarization_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|518.8 MB| + +## References + +https://huggingface.co/huyenquinn282/NDC_summarization_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_pipeline_en.md new file mode 100644 index 00000000000000..fdd2deec157a49 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ndc_summarization_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ndc_summarization_v1_pipeline pipeline T5Transformer from huyenquinn282 +author: John Snow Labs +name: ndc_summarization_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ndc_summarization_v1_pipeline` is a English model originally trained by huyenquinn282. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ndc_summarization_v1_pipeline_en_5.4.2_3.0_1723969592625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ndc_summarization_v1_pipeline_en_5.4.2_3.0_1723969592625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ndc_summarization_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ndc_summarization_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ndc_summarization_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|518.8 MB| + +## References + +https://huggingface.co/huyenquinn282/NDC_summarization_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_en.md b/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_en.md new file mode 100644 index 00000000000000..0f05bc12224408 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nlp_quizcraft_mcq T5Transformer from ZoroHunter +author: John Snow Labs +name: nlp_quizcraft_mcq +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nlp_quizcraft_mcq` is a English model originally trained by ZoroHunter. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nlp_quizcraft_mcq_en_5.4.2_3.0_1723995394068.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nlp_quizcraft_mcq_en_5.4.2_3.0_1723995394068.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nlp_quizcraft_mcq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nlp_quizcraft_mcq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nlp_quizcraft_mcq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ZoroHunter/NLP_quizcraft_mcq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_pipeline_en.md new file mode 100644 index 00000000000000..d98a76a65c555f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nlp_quizcraft_mcq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nlp_quizcraft_mcq_pipeline pipeline T5Transformer from ZoroHunter +author: John Snow Labs +name: nlp_quizcraft_mcq_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nlp_quizcraft_mcq_pipeline` is a English model originally trained by ZoroHunter. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nlp_quizcraft_mcq_pipeline_en_5.4.2_3.0_1723995439985.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nlp_quizcraft_mcq_pipeline_en_5.4.2_3.0_1723995439985.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nlp_quizcraft_mcq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nlp_quizcraft_mcq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nlp_quizcraft_mcq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ZoroHunter/NLP_quizcraft_mcq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_en.md new file mode 100644 index 00000000000000..914bef6273d92b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English normal_nodes_normal_graphs_with_edge_document_level_t5_run1 T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_normal_graphs_with_edge_document_level_t5_run1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_normal_graphs_with_edge_document_level_t5_run1` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run1_en_5.4.2_3.0_1723988633425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run1_en_5.4.2_3.0_1723988633425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("normal_nodes_normal_graphs_with_edge_document_level_t5_run1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("normal_nodes_normal_graphs_with_edge_document_level_t5_run1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_normal_graphs_with_edge_document_level_t5_run1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|311.0 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_normal_graphs_with_edge_document_level_T5_run1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline_en.md new file mode 100644 index 00000000000000..ad811920057361 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline_en_5.4.2_3.0_1723988653957.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline_en_5.4.2_3.0_1723988653957.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_normal_graphs_with_edge_document_level_t5_run1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|311.0 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_normal_graphs_with_edge_document_level_T5_run1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_en.md new file mode 100644 index 00000000000000..8a74647370995c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English normal_nodes_normal_graphs_with_edge_document_level_t5_run2 T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_normal_graphs_with_edge_document_level_t5_run2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_normal_graphs_with_edge_document_level_t5_run2` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run2_en_5.4.2_3.0_1723999374469.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run2_en_5.4.2_3.0_1723999374469.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("normal_nodes_normal_graphs_with_edge_document_level_t5_run2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("normal_nodes_normal_graphs_with_edge_document_level_t5_run2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_normal_graphs_with_edge_document_level_t5_run2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|311.1 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_normal_graphs_with_edge_document_level_T5_run2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline_en.md new file mode 100644 index 00000000000000..deeb0a554bafb8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723999392625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723999392625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_normal_graphs_with_edge_document_level_t5_run2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|311.1 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_normal_graphs_with_edge_document_level_T5_run2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md new file mode 100644 index 00000000000000..3be8559583c30c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2 T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1723995924702.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1723995924702.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|313.9 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_shuffled_graphs_without_edge_document_level_T5_run2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md new file mode 100644 index 00000000000000..8912f6afd11a9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723995942495.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723995942495.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|normal_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|313.9 MB| + +## References + +https://huggingface.co/sheoran95/normal_nodes_shuffled_graphs_without_edge_document_level_T5_run2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_en.md b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_en.md new file mode 100644 index 00000000000000..0dfd76baefd95f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nt5_small_drop T5Transformer from StonyBrookNLP +author: John Snow Labs +name: nt5_small_drop +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nt5_small_drop` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nt5_small_drop_en_5.4.2_3.0_1724001235408.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nt5_small_drop_en_5.4.2_3.0_1724001235408.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nt5_small_drop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nt5_small_drop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nt5_small_drop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/StonyBrookNLP/nt5-small-drop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_pipeline_en.md new file mode 100644 index 00000000000000..bc9a19d9b2aa98 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_drop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nt5_small_drop_pipeline pipeline T5Transformer from StonyBrookNLP +author: John Snow Labs +name: nt5_small_drop_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nt5_small_drop_pipeline` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nt5_small_drop_pipeline_en_5.4.2_3.0_1724001252316.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nt5_small_drop_pipeline_en_5.4.2_3.0_1724001252316.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nt5_small_drop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nt5_small_drop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nt5_small_drop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/StonyBrookNLP/nt5-small-drop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_en.md b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_en.md new file mode 100644 index 00000000000000..c904a84ce68380 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nt5_small_rc1 T5Transformer from nielsr +author: John Snow Labs +name: nt5_small_rc1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nt5_small_rc1` is a English model originally trained by nielsr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nt5_small_rc1_en_5.4.2_3.0_1724023573368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nt5_small_rc1_en_5.4.2_3.0_1724023573368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nt5_small_rc1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nt5_small_rc1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nt5_small_rc1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|347.8 MB| + +## References + +https://huggingface.co/nielsr/nt5-small-rc1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_pipeline_en.md new file mode 100644 index 00000000000000..7ba7a42b5a94ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-nt5_small_rc1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nt5_small_rc1_pipeline pipeline T5Transformer from nielsr +author: John Snow Labs +name: nt5_small_rc1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nt5_small_rc1_pipeline` is a English model originally trained by nielsr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nt5_small_rc1_pipeline_en_5.4.2_3.0_1724023589692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nt5_small_rc1_pipeline_en_5.4.2_3.0_1724023589692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nt5_small_rc1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nt5_small_rc1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nt5_small_rc1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|347.8 MB| + +## References + +https://huggingface.co/nielsr/nt5-small-rc1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_en.md b/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_en.md new file mode 100644 index 00000000000000..2b0ac12aa2cb81 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English oop_german_qg_flan_t5_base_v3 T5Transformer from LunaticTanuki +author: John Snow Labs +name: oop_german_qg_flan_t5_base_v3 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`oop_german_qg_flan_t5_base_v3` is a English model originally trained by LunaticTanuki. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/oop_german_qg_flan_t5_base_v3_en_5.4.2_3.0_1723948102578.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/oop_german_qg_flan_t5_base_v3_en_5.4.2_3.0_1723948102578.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("oop_german_qg_flan_t5_base_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("oop_german_qg_flan_t5_base_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|oop_german_qg_flan_t5_base_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LunaticTanuki/oop-de-qg-flan-t5-base-v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_pipeline_en.md new file mode 100644 index 00000000000000..d8368cd95c517b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-oop_german_qg_flan_t5_base_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English oop_german_qg_flan_t5_base_v3_pipeline pipeline T5Transformer from LunaticTanuki +author: John Snow Labs +name: oop_german_qg_flan_t5_base_v3_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`oop_german_qg_flan_t5_base_v3_pipeline` is a English model originally trained by LunaticTanuki. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/oop_german_qg_flan_t5_base_v3_pipeline_en_5.4.2_3.0_1723948151079.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/oop_german_qg_flan_t5_base_v3_pipeline_en_5.4.2_3.0_1723948151079.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("oop_german_qg_flan_t5_base_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("oop_german_qg_flan_t5_base_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|oop_german_qg_flan_t5_base_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LunaticTanuki/oop-de-qg-flan-t5-base-v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_en.md b/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_en.md new file mode 100644 index 00000000000000..bd252ba092926d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English peft_finetune_flan_t5_mc_question_generation T5Transformer from Michelvh +author: John Snow Labs +name: peft_finetune_flan_t5_mc_question_generation +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`peft_finetune_flan_t5_mc_question_generation` is a English model originally trained by Michelvh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/peft_finetune_flan_t5_mc_question_generation_en_5.4.2_3.0_1723994917756.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/peft_finetune_flan_t5_mc_question_generation_en_5.4.2_3.0_1723994917756.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("peft_finetune_flan_t5_mc_question_generation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("peft_finetune_flan_t5_mc_question_generation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|peft_finetune_flan_t5_mc_question_generation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|700.8 MB| + +## References + +https://huggingface.co/Michelvh/peft-finetune-flan-t5-mc-question-generation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_pipeline_en.md new file mode 100644 index 00000000000000..870885cfe373cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-peft_finetune_flan_t5_mc_question_generation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English peft_finetune_flan_t5_mc_question_generation_pipeline pipeline T5Transformer from Michelvh +author: John Snow Labs +name: peft_finetune_flan_t5_mc_question_generation_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`peft_finetune_flan_t5_mc_question_generation_pipeline` is a English model originally trained by Michelvh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/peft_finetune_flan_t5_mc_question_generation_pipeline_en_5.4.2_3.0_1723995085027.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/peft_finetune_flan_t5_mc_question_generation_pipeline_en_5.4.2_3.0_1723995085027.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("peft_finetune_flan_t5_mc_question_generation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("peft_finetune_flan_t5_mc_question_generation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|peft_finetune_flan_t5_mc_question_generation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|700.8 MB| + +## References + +https://huggingface.co/Michelvh/peft-finetune-flan-t5-mc-question-generation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_en.md b/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_en.md new file mode 100644 index 00000000000000..539f552c32cc94 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English peft_t5_base_50_fullshot T5Transformer from niceanyh +author: John Snow Labs +name: peft_t5_base_50_fullshot +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`peft_t5_base_50_fullshot` is a English model originally trained by niceanyh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/peft_t5_base_50_fullshot_en_5.4.2_3.0_1723983149247.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/peft_t5_base_50_fullshot_en_5.4.2_3.0_1723983149247.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("peft_t5_base_50_fullshot","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("peft_t5_base_50_fullshot", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|peft_t5_base_50_fullshot| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/niceanyh/peft-T5-base-50-fullshot \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_pipeline_en.md new file mode 100644 index 00000000000000..a3aaa2fe81e288 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-peft_t5_base_50_fullshot_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English peft_t5_base_50_fullshot_pipeline pipeline T5Transformer from niceanyh +author: John Snow Labs +name: peft_t5_base_50_fullshot_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`peft_t5_base_50_fullshot_pipeline` is a English model originally trained by niceanyh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/peft_t5_base_50_fullshot_pipeline_en_5.4.2_3.0_1723983197881.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/peft_t5_base_50_fullshot_pipeline_en_5.4.2_3.0_1723983197881.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("peft_t5_base_50_fullshot_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("peft_t5_base_50_fullshot_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|peft_t5_base_50_fullshot_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/niceanyh/peft-T5-base-50-fullshot + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_en.md new file mode 100644 index 00000000000000..232e8ceb1443ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English plainly_v1 T5Transformer from mynti +author: John Snow Labs +name: plainly_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`plainly_v1` is a English model originally trained by mynti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/plainly_v1_en_5.4.2_3.0_1724009802056.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/plainly_v1_en_5.4.2_3.0_1724009802056.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("plainly_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("plainly_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|plainly_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mynti/plainly-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_pipeline_en.md new file mode 100644 index 00000000000000..6829a3072f552f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-plainly_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English plainly_v1_pipeline pipeline T5Transformer from mynti +author: John Snow Labs +name: plainly_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`plainly_v1_pipeline` is a English model originally trained by mynti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/plainly_v1_pipeline_en_5.4.2_3.0_1724009852793.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/plainly_v1_pipeline_en_5.4.2_3.0_1724009852793.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("plainly_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("plainly_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|plainly_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mynti/plainly-v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_en.md b/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_en.md new file mode 100644 index 00000000000000..86fc1ffb687a65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English poleval2021_task4_plt5_base_qa T5Transformer from piotr-rybak +author: John Snow Labs +name: poleval2021_task4_plt5_base_qa +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poleval2021_task4_plt5_base_qa` is a English model originally trained by piotr-rybak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poleval2021_task4_plt5_base_qa_en_5.4.2_3.0_1723947720988.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poleval2021_task4_plt5_base_qa_en_5.4.2_3.0_1723947720988.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("poleval2021_task4_plt5_base_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("poleval2021_task4_plt5_base_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poleval2021_task4_plt5_base_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/piotr-rybak/poleval2021-task4-plt5-base-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_pipeline_en.md new file mode 100644 index 00000000000000..2f74ccb2afc43e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-poleval2021_task4_plt5_base_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English poleval2021_task4_plt5_base_qa_pipeline pipeline T5Transformer from piotr-rybak +author: John Snow Labs +name: poleval2021_task4_plt5_base_qa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poleval2021_task4_plt5_base_qa_pipeline` is a English model originally trained by piotr-rybak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poleval2021_task4_plt5_base_qa_pipeline_en_5.4.2_3.0_1723947783460.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poleval2021_task4_plt5_base_qa_pipeline_en_5.4.2_3.0_1723947783460.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("poleval2021_task4_plt5_base_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("poleval2021_task4_plt5_base_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poleval2021_task4_plt5_base_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/piotr-rybak/poleval2021-task4-plt5-base-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_en.md new file mode 100644 index 00000000000000..42529347a866d8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English polish_transliterator_small T5Transformer from marcus2000 +author: John Snow Labs +name: polish_transliterator_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polish_transliterator_small` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polish_transliterator_small_en_5.4.2_3.0_1724000974388.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polish_transliterator_small_en_5.4.2_3.0_1724000974388.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("polish_transliterator_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("polish_transliterator_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polish_transliterator_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|270.1 MB| + +## References + +https://huggingface.co/marcus2000/polish_transliterator_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_pipeline_en.md new file mode 100644 index 00000000000000..9141d0fb9b456b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-polish_transliterator_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English polish_transliterator_small_pipeline pipeline T5Transformer from marcus2000 +author: John Snow Labs +name: polish_transliterator_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polish_transliterator_small_pipeline` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polish_transliterator_small_pipeline_en_5.4.2_3.0_1724001009782.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polish_transliterator_small_pipeline_en_5.4.2_3.0_1724001009782.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("polish_transliterator_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("polish_transliterator_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polish_transliterator_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|270.1 MB| + +## References + +https://huggingface.co/marcus2000/polish_transliterator_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-predy_en.md b/docs/_posts/ahmedlone127/2024-08-18-predy_en.md new file mode 100644 index 00000000000000..2d24790c796449 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-predy_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English predy T5Transformer from Ghani-25 +author: John Snow Labs +name: predy +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`predy` is a English model originally trained by Ghani-25. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/predy_en_5.4.2_3.0_1723940289756.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/predy_en_5.4.2_3.0_1723940289756.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("predy","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("predy", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|predy| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|962.2 MB| + +## References + +https://huggingface.co/Ghani-25/predy \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-predy_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-predy_pipeline_en.md new file mode 100644 index 00000000000000..a1cacb521547b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-predy_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English predy_pipeline pipeline T5Transformer from Ghani-25 +author: John Snow Labs +name: predy_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`predy_pipeline` is a English model originally trained by Ghani-25. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/predy_pipeline_en_5.4.2_3.0_1723940347200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/predy_pipeline_en_5.4.2_3.0_1723940347200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("predy_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("predy_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|predy_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|962.2 MB| + +## References + +https://huggingface.co/Ghani-25/predy + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_pipeline_zh.md b/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_pipeline_zh.md new file mode 100644 index 00000000000000..91722a59ae2210 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_pipeline_zh.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Chinese promptclue_base_pipeline pipeline T5Transformer from ClueAI +author: John Snow Labs +name: promptclue_base_pipeline +date: 2024-08-18 +tags: [zh, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`promptclue_base_pipeline` is a Chinese model originally trained by ClueAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/promptclue_base_pipeline_zh_5.4.2_3.0_1724014787681.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/promptclue_base_pipeline_zh_5.4.2_3.0_1724014787681.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("promptclue_base_pipeline", lang = "zh") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("promptclue_base_pipeline", lang = "zh") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|promptclue_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|zh| +|Size:|520.9 MB| + +## References + +https://huggingface.co/ClueAI/PromptCLUE-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_zh.md b/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_zh.md new file mode 100644 index 00000000000000..2f0606b445255c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-promptclue_base_zh.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Chinese promptclue_base T5Transformer from ClueAI +author: John Snow Labs +name: promptclue_base +date: 2024-08-18 +tags: [zh, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`promptclue_base` is a Chinese model originally trained by ClueAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/promptclue_base_zh_5.4.2_3.0_1724014619407.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/promptclue_base_zh_5.4.2_3.0_1724014619407.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("promptclue_base","zh") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("promptclue_base", "zh") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|promptclue_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|zh| +|Size:|520.9 MB| + +## References + +https://huggingface.co/ClueAI/PromptCLUE-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pipeline_pt.md b/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pipeline_pt.md new file mode 100644 index 00000000000000..10fc64fbbf6dd6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pipeline_pt.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Portuguese ptt5_small_portuguese_vocab_pipeline pipeline T5Transformer from unicamp-dl +author: John Snow Labs +name: ptt5_small_portuguese_vocab_pipeline +date: 2024-08-18 +tags: [pt, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_small_portuguese_vocab_pipeline` is a Portuguese model originally trained by unicamp-dl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_small_portuguese_vocab_pipeline_pt_5.4.2_3.0_1724012537337.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_small_portuguese_vocab_pipeline_pt_5.4.2_3.0_1724012537337.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ptt5_small_portuguese_vocab_pipeline", lang = "pt") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ptt5_small_portuguese_vocab_pipeline", lang = "pt") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_small_portuguese_vocab_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|pt| +|Size:|178.8 MB| + +## References + +https://huggingface.co/unicamp-dl/ptt5-small-portuguese-vocab + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pt.md b/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pt.md new file mode 100644 index 00000000000000..4e9b7621c045e0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ptt5_small_portuguese_vocab_pt.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Portuguese ptt5_small_portuguese_vocab T5Transformer from unicamp-dl +author: John Snow Labs +name: ptt5_small_portuguese_vocab +date: 2024-08-18 +tags: [pt, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_small_portuguese_vocab` is a Portuguese model originally trained by unicamp-dl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_small_portuguese_vocab_pt_5.4.2_3.0_1724012479512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_small_portuguese_vocab_pt_5.4.2_3.0_1724012479512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ptt5_small_portuguese_vocab","pt") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ptt5_small_portuguese_vocab", "pt") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_small_portuguese_vocab| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|pt| +|Size:|178.8 MB| + +## References + +https://huggingface.co/unicamp-dl/ptt5-small-portuguese-vocab \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_en.md b/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_en.md new file mode 100644 index 00000000000000..77a25482d15881 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ptt5_wikilingua_1024 T5Transformer from arthurmluz +author: John Snow Labs +name: ptt5_wikilingua_1024 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_wikilingua_1024` is a English model originally trained by arthurmluz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_wikilingua_1024_en_5.4.2_3.0_1724017243345.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_wikilingua_1024_en_5.4.2_3.0_1724017243345.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ptt5_wikilingua_1024","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ptt5_wikilingua_1024", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_wikilingua_1024| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|998.9 MB| + +## References + +https://huggingface.co/arthurmluz/ptt5-wikilingua-1024 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_pipeline_en.md new file mode 100644 index 00000000000000..a1b1e03bef74ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ptt5_wikilingua_1024_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ptt5_wikilingua_1024_pipeline pipeline T5Transformer from arthurmluz +author: John Snow Labs +name: ptt5_wikilingua_1024_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_wikilingua_1024_pipeline` is a English model originally trained by arthurmluz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_wikilingua_1024_pipeline_en_5.4.2_3.0_1724017291545.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_wikilingua_1024_pipeline_en_5.4.2_3.0_1724017291545.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ptt5_wikilingua_1024_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ptt5_wikilingua_1024_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_wikilingua_1024_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|998.9 MB| + +## References + +https://huggingface.co/arthurmluz/ptt5-wikilingua-1024 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_en.md b/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_en.md new file mode 100644 index 00000000000000..207ba3c7f52a41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English punctuation_tedtalk2012_t5_large T5Transformer from tiagoblima +author: John Snow Labs +name: punctuation_tedtalk2012_t5_large +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`punctuation_tedtalk2012_t5_large` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/punctuation_tedtalk2012_t5_large_en_5.4.2_3.0_1724024301799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/punctuation_tedtalk2012_t5_large_en_5.4.2_3.0_1724024301799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("punctuation_tedtalk2012_t5_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("punctuation_tedtalk2012_t5_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|punctuation_tedtalk2012_t5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/tiagoblima/punctuation-tedtalk2012-t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_pipeline_en.md new file mode 100644 index 00000000000000..515b48e016de58 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-punctuation_tedtalk2012_t5_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English punctuation_tedtalk2012_t5_large_pipeline pipeline T5Transformer from tiagoblima +author: John Snow Labs +name: punctuation_tedtalk2012_t5_large_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`punctuation_tedtalk2012_t5_large_pipeline` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/punctuation_tedtalk2012_t5_large_pipeline_en_5.4.2_3.0_1724024446368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/punctuation_tedtalk2012_t5_large_pipeline_en_5.4.2_3.0_1724024446368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("punctuation_tedtalk2012_t5_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("punctuation_tedtalk2012_t5_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|punctuation_tedtalk2012_t5_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/tiagoblima/punctuation-tedtalk2012-t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_en.md b/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_en.md new file mode 100644 index 00000000000000..15a753d49e55f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qcpg_questions T5Transformer from ibm +author: John Snow Labs +name: qcpg_questions +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_questions` is a English model originally trained by ibm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_questions_en_5.4.2_3.0_1724022716139.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_questions_en_5.4.2_3.0_1724022716139.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qcpg_questions","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qcpg_questions", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_questions| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ibm/qcpg-questions \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_pipeline_en.md new file mode 100644 index 00000000000000..3d1ac456a34ead --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-qcpg_questions_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qcpg_questions_pipeline pipeline T5Transformer from ibm +author: John Snow Labs +name: qcpg_questions_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_questions_pipeline` is a English model originally trained by ibm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_questions_pipeline_en_5.4.2_3.0_1724022762286.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_questions_pipeline_en_5.4.2_3.0_1724022762286.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qcpg_questions_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qcpg_questions_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_questions_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ibm/qcpg-questions + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_en.md b/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_en.md new file mode 100644 index 00000000000000..e583d32ee26a8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qg_shredder_31 T5Transformer from shredder-31 +author: John Snow Labs +name: qg_shredder_31 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qg_shredder_31` is a English model originally trained by shredder-31. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qg_shredder_31_en_5.4.2_3.0_1724011813579.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qg_shredder_31_en_5.4.2_3.0_1724011813579.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qg_shredder_31","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qg_shredder_31", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qg_shredder_31| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/shredder-31/QG \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_pipeline_en.md new file mode 100644 index 00000000000000..6bb5d68ab23577 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-qg_shredder_31_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qg_shredder_31_pipeline pipeline T5Transformer from shredder-31 +author: John Snow Labs +name: qg_shredder_31_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qg_shredder_31_pipeline` is a English model originally trained by shredder-31. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qg_shredder_31_pipeline_en_5.4.2_3.0_1724011963085.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qg_shredder_31_pipeline_en_5.4.2_3.0_1724011963085.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qg_shredder_31_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qg_shredder_31_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qg_shredder_31_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/shredder-31/QG + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_en.md b/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_en.md new file mode 100644 index 00000000000000..d943c9002c7738 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English query_decision_train_on_maybe_valid T5Transformer from JackBAI +author: John Snow Labs +name: query_decision_train_on_maybe_valid +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`query_decision_train_on_maybe_valid` is a English model originally trained by JackBAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/query_decision_train_on_maybe_valid_en_5.4.2_3.0_1723946046853.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/query_decision_train_on_maybe_valid_en_5.4.2_3.0_1723946046853.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("query_decision_train_on_maybe_valid","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("query_decision_train_on_maybe_valid", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|query_decision_train_on_maybe_valid| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/JackBAI/query_decision_train_on_maybe_valid \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_pipeline_en.md new file mode 100644 index 00000000000000..29d1013d030d42 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-query_decision_train_on_maybe_valid_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English query_decision_train_on_maybe_valid_pipeline pipeline T5Transformer from JackBAI +author: John Snow Labs +name: query_decision_train_on_maybe_valid_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`query_decision_train_on_maybe_valid_pipeline` is a English model originally trained by JackBAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/query_decision_train_on_maybe_valid_pipeline_en_5.4.2_3.0_1723946096129.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/query_decision_train_on_maybe_valid_pipeline_en_5.4.2_3.0_1723946096129.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("query_decision_train_on_maybe_valid_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("query_decision_train_on_maybe_valid_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|query_decision_train_on_maybe_valid_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/JackBAI/query_decision_train_on_maybe_valid + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pipeline_pt.md b/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pipeline_pt.md new file mode 100644 index 00000000000000..fc15b405623a4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pipeline_pt.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Portuguese question_generation_t5_base_portuguese_breton_pipeline pipeline T5Transformer from vabatista +author: John Snow Labs +name: question_generation_t5_base_portuguese_breton_pipeline +date: 2024-08-18 +tags: [pt, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_generation_t5_base_portuguese_breton_pipeline` is a Portuguese model originally trained by vabatista. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_generation_t5_base_portuguese_breton_pipeline_pt_5.4.2_3.0_1724021208588.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_generation_t5_base_portuguese_breton_pipeline_pt_5.4.2_3.0_1724021208588.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("question_generation_t5_base_portuguese_breton_pipeline", lang = "pt") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("question_generation_t5_base_portuguese_breton_pipeline", lang = "pt") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_generation_t5_base_portuguese_breton_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|pt| +|Size:|1.0 GB| + +## References + +https://huggingface.co/vabatista/question-generation-t5-base-pt-br + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pt.md b/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pt.md new file mode 100644 index 00000000000000..f5dbec545e0cf9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-question_generation_t5_base_portuguese_breton_pt.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Portuguese question_generation_t5_base_portuguese_breton T5Transformer from vabatista +author: John Snow Labs +name: question_generation_t5_base_portuguese_breton +date: 2024-08-18 +tags: [pt, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: pt +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_generation_t5_base_portuguese_breton` is a Portuguese model originally trained by vabatista. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_generation_t5_base_portuguese_breton_pt_5.4.2_3.0_1724021160611.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_generation_t5_base_portuguese_breton_pt_5.4.2_3.0_1724021160611.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("question_generation_t5_base_portuguese_breton","pt") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("question_generation_t5_base_portuguese_breton", "pt") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_generation_t5_base_portuguese_breton| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|pt| +|Size:|1.0 GB| + +## References + +https://huggingface.co/vabatista/question-generation-t5-base-pt-br \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_en.md new file mode 100644 index 00000000000000..4c23152ad7b555 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English real_prompt_300_problem_gen_t5_small T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_problem_gen_t5_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_problem_gen_t5_small` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_problem_gen_t5_small_en_5.4.2_3.0_1723989699955.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_problem_gen_t5_small_en_5.4.2_3.0_1723989699955.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("real_prompt_300_problem_gen_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("real_prompt_300_problem_gen_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_problem_gen_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-problem-gen-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..c87f409f2b72f1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-real_prompt_300_problem_gen_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English real_prompt_300_problem_gen_t5_small_pipeline pipeline T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_300_problem_gen_t5_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_300_problem_gen_t5_small_pipeline` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_300_problem_gen_t5_small_pipeline_en_5.4.2_3.0_1723989758633.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_300_problem_gen_t5_small_pipeline_en_5.4.2_3.0_1723989758633.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("real_prompt_300_problem_gen_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("real_prompt_300_problem_gen_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_300_problem_gen_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-300-problem-gen-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-results_t5_en.md b/docs/_posts/ahmedlone127/2024-08-18-results_t5_en.md new file mode 100644 index 00000000000000..2e7813f4449c18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-results_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English results_t5 T5Transformer from mohsin-aslam +author: John Snow Labs +name: results_t5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_t5` is a English model originally trained by mohsin-aslam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_t5_en_5.4.2_3.0_1723960217799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_t5_en_5.4.2_3.0_1723960217799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("results_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("results_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/mohsin-aslam/results_t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-results_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-results_t5_pipeline_en.md new file mode 100644 index 00000000000000..e956231ce58517 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-results_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English results_t5_pipeline pipeline T5Transformer from mohsin-aslam +author: John Snow Labs +name: results_t5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_t5_pipeline` is a English model originally trained by mohsin-aslam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_t5_pipeline_en_5.4.2_3.0_1723960392944.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_t5_pipeline_en_5.4.2_3.0_1723960392944.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("results_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("results_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/mohsin-aslam/results_t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_en.md b/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_en.md new file mode 100644 index 00000000000000..18d28f28180bd9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rut5_base_simple_aphasia T5Transformer from ShinjuM +author: John Snow Labs +name: rut5_base_simple_aphasia +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_simple_aphasia` is a English model originally trained by ShinjuM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_simple_aphasia_en_5.4.2_3.0_1723991842806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_simple_aphasia_en_5.4.2_3.0_1723991842806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_base_simple_aphasia","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_base_simple_aphasia", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_simple_aphasia| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.7 MB| + +## References + +https://huggingface.co/ShinjuM/rut5-base-simple-aphasia \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_pipeline_en.md new file mode 100644 index 00000000000000..20df7a2a70712f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-rut5_base_simple_aphasia_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rut5_base_simple_aphasia_pipeline pipeline T5Transformer from ShinjuM +author: John Snow Labs +name: rut5_base_simple_aphasia_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_simple_aphasia_pipeline` is a English model originally trained by ShinjuM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_simple_aphasia_pipeline_en_5.4.2_3.0_1723991887357.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_simple_aphasia_pipeline_en_5.4.2_3.0_1723991887357.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_base_simple_aphasia_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_base_simple_aphasia_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_simple_aphasia_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.7 MB| + +## References + +https://huggingface.co/ShinjuM/rut5-base-simple-aphasia + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_en.md b/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_en.md new file mode 100644 index 00000000000000..7da28d61ed3b8e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English shreyas_t5_finetuned_samsum T5Transformer from shreyas1104 +author: John Snow Labs +name: shreyas_t5_finetuned_samsum +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shreyas_t5_finetuned_samsum` is a English model originally trained by shreyas1104. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shreyas_t5_finetuned_samsum_en_5.4.2_3.0_1723990993474.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shreyas_t5_finetuned_samsum_en_5.4.2_3.0_1723990993474.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("shreyas_t5_finetuned_samsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("shreyas_t5_finetuned_samsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shreyas_t5_finetuned_samsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.3 MB| + +## References + +https://huggingface.co/shreyas1104/shreyas_t5_finetuned_samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_pipeline_en.md new file mode 100644 index 00000000000000..605c91d661335a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-shreyas_t5_finetuned_samsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English shreyas_t5_finetuned_samsum_pipeline pipeline T5Transformer from shreyas1104 +author: John Snow Labs +name: shreyas_t5_finetuned_samsum_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shreyas_t5_finetuned_samsum_pipeline` is a English model originally trained by shreyas1104. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shreyas_t5_finetuned_samsum_pipeline_en_5.4.2_3.0_1723991010421.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shreyas_t5_finetuned_samsum_pipeline_en_5.4.2_3.0_1723991010421.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("shreyas_t5_finetuned_samsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("shreyas_t5_finetuned_samsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shreyas_t5_finetuned_samsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.3 MB| + +## References + +https://huggingface.co/shreyas1104/shreyas_t5_finetuned_samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md b/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md new file mode 100644 index 00000000000000..3e0dab15972e76 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2 T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1723994312452.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1723994312452.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|314.0 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_shuffled_graphs_without_edge_document_level_T5_run2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md new file mode 100644 index 00000000000000..5405ee4808130f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723994330681.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1723994330681.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_shuffled_graphs_without_edge_document_level_t5_run2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|314.0 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_shuffled_graphs_without_edge_document_level_T5_run2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_en.md b/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_en.md new file mode 100644 index 00000000000000..cba98f0d094ad3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sno_empty2_5k_msp T5Transformer from Roy029 +author: John Snow Labs +name: sno_empty2_5k_msp +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sno_empty2_5k_msp` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sno_empty2_5k_msp_en_5.4.2_3.0_1723957970368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sno_empty2_5k_msp_en_5.4.2_3.0_1723957970368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sno_empty2_5k_msp","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sno_empty2_5k_msp", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sno_empty2_5k_msp| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|941.0 MB| + +## References + +https://huggingface.co/Roy029/sno_empty2_5k_msp \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_pipeline_en.md new file mode 100644 index 00000000000000..1634ccf2e50443 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sno_empty2_5k_msp_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sno_empty2_5k_msp_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: sno_empty2_5k_msp_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sno_empty2_5k_msp_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sno_empty2_5k_msp_pipeline_en_5.4.2_3.0_1723958039855.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sno_empty2_5k_msp_pipeline_en_5.4.2_3.0_1723958039855.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sno_empty2_5k_msp_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sno_empty2_5k_msp_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sno_empty2_5k_msp_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|941.0 MB| + +## References + +https://huggingface.co/Roy029/sno_empty2_5k_msp + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_en.md b/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_en.md new file mode 100644 index 00000000000000..95ab6261aa7fac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sno_extend_py5k T5Transformer from Roy029 +author: John Snow Labs +name: sno_extend_py5k +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sno_extend_py5k` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sno_extend_py5k_en_5.4.2_3.0_1723996348326.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sno_extend_py5k_en_5.4.2_3.0_1723996348326.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sno_extend_py5k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sno_extend_py5k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sno_extend_py5k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|565.4 MB| + +## References + +https://huggingface.co/Roy029/sno_extend_py5k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_pipeline_en.md new file mode 100644 index 00000000000000..96a54aa0296660 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sno_extend_py5k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sno_extend_py5k_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: sno_extend_py5k_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sno_extend_py5k_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sno_extend_py5k_pipeline_en_5.4.2_3.0_1723996517264.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sno_extend_py5k_pipeline_en_5.4.2_3.0_1723996517264.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sno_extend_py5k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sno_extend_py5k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sno_extend_py5k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|565.4 MB| + +## References + +https://huggingface.co/Roy029/sno_extend_py5k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_en.md b/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_en.md new file mode 100644 index 00000000000000..545137371614d8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English speller_example_summervent T5Transformer from summervent +author: John Snow Labs +name: speller_example_summervent +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`speller_example_summervent` is a English model originally trained by summervent. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/speller_example_summervent_en_5.4.2_3.0_1723965285199.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/speller_example_summervent_en_5.4.2_3.0_1723965285199.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("speller_example_summervent","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("speller_example_summervent", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|speller_example_summervent| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/summervent/speller-example \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_pipeline_en.md new file mode 100644 index 00000000000000..1ad2ad30bb9449 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-speller_example_summervent_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English speller_example_summervent_pipeline pipeline T5Transformer from summervent +author: John Snow Labs +name: speller_example_summervent_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`speller_example_summervent_pipeline` is a English model originally trained by summervent. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/speller_example_summervent_pipeline_en_5.4.2_3.0_1723965334281.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/speller_example_summervent_pipeline_en_5.4.2_3.0_1723965334281.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("speller_example_summervent_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("speller_example_summervent_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|speller_example_summervent_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/summervent/speller-example + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sum_italian_en.md b/docs/_posts/ahmedlone127/2024-08-18-sum_italian_en.md new file mode 100644 index 00000000000000..353c533b0e4425 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sum_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sum_italian T5Transformer from CoderCoy +author: John Snow Labs +name: sum_italian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_italian` is a English model originally trained by CoderCoy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_italian_en_5.4.2_3.0_1723989965164.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_italian_en_5.4.2_3.0_1723989965164.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sum_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sum_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|317.1 MB| + +## References + +https://huggingface.co/CoderCoy/sum_it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sum_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-sum_italian_pipeline_en.md new file mode 100644 index 00000000000000..7a088d297e5888 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sum_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sum_italian_pipeline pipeline T5Transformer from CoderCoy +author: John Snow Labs +name: sum_italian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_italian_pipeline` is a English model originally trained by CoderCoy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_italian_pipeline_en_5.4.2_3.0_1723989988428.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_italian_pipeline_en_5.4.2_3.0_1723989988428.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sum_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sum_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|317.1 MB| + +## References + +https://huggingface.co/CoderCoy/sum_it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sum_sum_en.md b/docs/_posts/ahmedlone127/2024-08-18-sum_sum_en.md new file mode 100644 index 00000000000000..7eac257597150b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sum_sum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sum_sum T5Transformer from SamuelMiller +author: John Snow Labs +name: sum_sum +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_sum` is a English model originally trained by SamuelMiller. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_sum_en_5.4.2_3.0_1723965193569.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_sum_en_5.4.2_3.0_1723965193569.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sum_sum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sum_sum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_sum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/SamuelMiller/sum_sum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-sum_sum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-sum_sum_pipeline_en.md new file mode 100644 index 00000000000000..1886bf14bea448 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-sum_sum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sum_sum_pipeline pipeline T5Transformer from SamuelMiller +author: John Snow Labs +name: sum_sum_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_sum_pipeline` is a English model originally trained by SamuelMiller. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_sum_pipeline_en_5.4.2_3.0_1723965210496.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_sum_pipeline_en_5.4.2_3.0_1723965210496.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sum_sum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sum_sum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_sum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/SamuelMiller/sum_sum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_en.md b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_en.md new file mode 100644 index 00000000000000..7507393f646cb2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarizer_google_long_t5_local_base_keybert_unfaceted T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_local_base_keybert_unfaceted +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_local_base_keybert_unfaceted` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_keybert_unfaceted_en_5.4.2_3.0_1723998835330.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_keybert_unfaceted_en_5.4.2_3.0_1723998835330.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarizer_google_long_t5_local_base_keybert_unfaceted","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarizer_google_long_t5_local_base_keybert_unfaceted", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_local_base_keybert_unfaceted| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-local-base_keybert_unfaceted \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline_en.md new file mode 100644 index 00000000000000..80f1b138751dd1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline pipeline T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline_en_5.4.2_3.0_1723998881116.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline_en_5.4.2_3.0_1723998881116.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_local_base_keybert_unfaceted_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-local-base_keybert_unfaceted + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_en.md b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_en.md new file mode 100644 index 00000000000000..c9dc4fb339f909 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarizer_google_long_t5_local_base_tf_idf_faceted T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_local_base_tf_idf_faceted +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_local_base_tf_idf_faceted` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_tf_idf_faceted_en_5.4.2_3.0_1723986937856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_tf_idf_faceted_en_5.4.2_3.0_1723986937856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarizer_google_long_t5_local_base_tf_idf_faceted","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarizer_google_long_t5_local_base_tf_idf_faceted", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_local_base_tf_idf_faceted| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-local-base_tf_idf_faceted \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline_en.md new file mode 100644 index 00000000000000..9d6442849e9e6c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline pipeline T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline_en_5.4.2_3.0_1723986986959.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline_en_5.4.2_3.0_1723986986959.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_local_base_tf_idf_faceted_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-local-base_tf_idf_faceted + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_en.md b/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_en.md new file mode 100644 index 00000000000000..fe7d2275866b43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English superglue_boolq T5Transformer from ShengdingHu +author: John Snow Labs +name: superglue_boolq +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`superglue_boolq` is a English model originally trained by ShengdingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/superglue_boolq_en_5.4.2_3.0_1724021952875.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/superglue_boolq_en_5.4.2_3.0_1724021952875.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("superglue_boolq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("superglue_boolq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|superglue_boolq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ShengdingHu/superglue-boolq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_pipeline_en.md new file mode 100644 index 00000000000000..f6c591e72e98fe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-superglue_boolq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English superglue_boolq_pipeline pipeline T5Transformer from ShengdingHu +author: John Snow Labs +name: superglue_boolq_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`superglue_boolq_pipeline` is a English model originally trained by ShengdingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/superglue_boolq_pipeline_en_5.4.2_3.0_1724022006581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/superglue_boolq_pipeline_en_5.4.2_3.0_1724022006581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("superglue_boolq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("superglue_boolq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|superglue_boolq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ShengdingHu/superglue-boolq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_en.md new file mode 100644 index 00000000000000..5ad9f14052c982 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_twitter_2016 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_twitter_2016 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_twitter_2016` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_twitter_2016_en_5.4.2_3.0_1724023152197.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_twitter_2016_en_5.4.2_3.0_1724023152197.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_twitter_2016","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_twitter_2016", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_twitter_2016| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-twitter-2016 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_pipeline_en.md new file mode 100644 index 00000000000000..da48e09a7994b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_twitter_2016_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_twitter_2016_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_twitter_2016_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_twitter_2016_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_twitter_2016_pipeline_en_5.4.2_3.0_1724023167907.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_twitter_2016_pipeline_en_5.4.2_3.0_1724023167907.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_twitter_2016_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_twitter_2016_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_twitter_2016_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-twitter-2016 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_en.md new file mode 100644 index 00000000000000..84a0dfe9705a24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2012_tonga_tonga_islands_2016 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2012_tonga_tonga_islands_2016 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2012_tonga_tonga_islands_2016` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_en_5.4.2_3.0_1723943433693.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_en_5.4.2_3.0_1723943433693.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2012_tonga_tonga_islands_2016","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2012_tonga_tonga_islands_2016", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2012_tonga_tonga_islands_2016| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2012_to_2016 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline_en.md new file mode 100644 index 00000000000000..22a6ac10797c0e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline_en_5.4.2_3.0_1723943450721.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline_en_5.4.2_3.0_1723943450721.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2012_tonga_tonga_islands_2016_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2012_to_2016 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_en.md new file mode 100644 index 00000000000000..98085a28d608e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2013_5 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2013_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2013_5` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_5_en_5.4.2_3.0_1723940194368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_5_en_5.4.2_3.0_1723940194368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2013_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2013_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2013_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2013-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_pipeline_en.md new file mode 100644 index 00000000000000..6d8faccd32e11f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2013_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2013_5_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2013_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2013_5_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_5_pipeline_en_5.4.2_3.0_1723940211216.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_5_pipeline_en_5.4.2_3.0_1723940211216.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2013_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2013_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2013_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2013-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_en.md new file mode 100644 index 00000000000000..4d60a69eb6dc70 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2017_8 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2017_8 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2017_8` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_8_en_5.4.2_3.0_1723961313061.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_8_en_5.4.2_3.0_1723961313061.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2017_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2017_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2017_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2017-8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_pipeline_en.md new file mode 100644 index 00000000000000..2b43c9f8f44c67 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2017_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2017_8_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2017_8_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2017_8_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_8_pipeline_en_5.4.2_3.0_1723961329867.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_8_pipeline_en_5.4.2_3.0_1723961329867.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2017_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2017_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2017_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2017-8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_en.md new file mode 100644 index 00000000000000..118fc24890fc5b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_8 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_8 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_8` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_8_en_5.4.2_3.0_1723966434064.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_8_en_5.4.2_3.0_1723966434064.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_pipeline_en.md new file mode 100644 index 00000000000000..0d8147e5ad2ad0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2018_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_8_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_8_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_8_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_8_pipeline_en_5.4.2_3.0_1723966450851.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_8_pipeline_en_5.4.2_3.0_1723966450851.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2018_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2018_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_en.md new file mode 100644 index 00000000000000..64b029108d2e3e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2021_11 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2021_11 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2021_11` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2021_11_en_5.4.2_3.0_1723950759410.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2021_11_en_5.4.2_3.0_1723950759410.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2021_11","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2021_11", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2021_11| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2021-11 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_pipeline_en.md new file mode 100644 index 00000000000000..1d5bebb06469ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_lm_wmt_2021_11_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2021_11_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2021_11_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2021_11_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2021_11_pipeline_en_5.4.2_3.0_1723950776680.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2021_11_pipeline_en_5.4.2_3.0_1723950776680.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2021_11_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2021_11_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2021_11_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2021-11 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_en.md new file mode 100644 index 00000000000000..d9be087d1b38d4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_0 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_0` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_0_en_5.4.2_3.0_1723995294047.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_0_en_5.4.2_3.0_1723995294047.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|298.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_pipeline_en.md new file mode 100644 index 00000000000000..a619f2b3b9dd79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2016_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_0_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_0_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_0_pipeline_en_5.4.2_3.0_1723995322153.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_0_pipeline_en_5.4.2_3.0_1723995322153.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2016_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2016_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|298.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_en.md new file mode 100644 index 00000000000000..bde463d0f5c370 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_10 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_10 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_10` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_10_en_5.4.2_3.0_1723990860915.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_10_en_5.4.2_3.0_1723990860915.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|300.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_pipeline_en.md new file mode 100644 index 00000000000000..2bf63d3e5631ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2017_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_10_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_10_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_10_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_10_pipeline_en_5.4.2_3.0_1723990888585.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_10_pipeline_en_5.4.2_3.0_1723990888585.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2017_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2017_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|300.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_en.md new file mode 100644 index 00000000000000..786adf31078ef7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2019_7 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2019_7 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2019_7` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_7_en_5.4.2_3.0_1723998401956.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_7_en_5.4.2_3.0_1723998401956.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2019_7","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2019_7", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2019_7| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|299.5 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2019-7 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_pipeline_en.md new file mode 100644 index 00000000000000..6539d1a91125ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2019_7_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2019_7_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2019_7_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2019_7_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_7_pipeline_en_5.4.2_3.0_1723998430248.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_7_pipeline_en_5.4.2_3.0_1723998430248.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2019_7_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2019_7_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2019_7_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|299.5 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2019-7 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_en.md new file mode 100644 index 00000000000000..c96a56077c8f06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_5 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_5` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_5_en_5.4.2_3.0_1723996876859.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_5_en_5.4.2_3.0_1723996876859.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|301.7 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_pipeline_en.md new file mode 100644 index 00000000000000..d12b040b414ed5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_5_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_5_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_5_pipeline_en_5.4.2_3.0_1723996903998.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_5_pipeline_en_5.4.2_3.0_1723996903998.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2020_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2020_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|301.7 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_en.md new file mode 100644 index 00000000000000..0b34af84d67e07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_7 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_7 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_7` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_7_en_5.4.2_3.0_1723940374725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_7_en_5.4.2_3.0_1723940374725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_7","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_7", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_7| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|300.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-7 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_pipeline_en.md new file mode 100644 index 00000000000000..9612d6b81d1c68 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_60m_poli_aff_2020_7_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_7_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_7_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_7_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_7_pipeline_en_5.4.2_3.0_1723940403581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_7_pipeline_en_5.4.2_3.0_1723940403581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2020_7_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2020_7_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_7_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|300.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-7 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_6m_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_6m_en.md new file mode 100644 index 00000000000000..8190a07242e231 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_6m_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_6m T5Transformer from versae +author: John Snow Labs +name: t5_6m +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_6m` is a English model originally trained by versae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_6m_en_5.4.2_3.0_1723979212205.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_6m_en_5.4.2_3.0_1723979212205.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_6m","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_6m", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_6m| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/versae/t5-6m \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_6m_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_6m_pipeline_en.md new file mode 100644 index 00000000000000..c93d63c313d112 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_6m_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_6m_pipeline pipeline T5Transformer from versae +author: John Snow Labs +name: t5_6m_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_6m_pipeline` is a English model originally trained by versae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_6m_pipeline_en_5.4.2_3.0_1723979259251.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_6m_pipeline_en_5.4.2_3.0_1723979259251.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_6m_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_6m_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_6m_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/versae/t5-6m + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_en.md new file mode 100644 index 00000000000000..10a9c7bf9a7fe0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_abs_qa T5Transformer from tuner007 +author: John Snow Labs +name: t5_abs_qa +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_abs_qa` is a English model originally trained by tuner007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_abs_qa_en_5.4.2_3.0_1724014949062.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_abs_qa_en_5.4.2_3.0_1724014949062.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_abs_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_abs_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_abs_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|988.8 MB| + +## References + +https://huggingface.co/tuner007/t5_abs_qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_pipeline_en.md new file mode 100644 index 00000000000000..a368b202196859 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_abs_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_abs_qa_pipeline pipeline T5Transformer from tuner007 +author: John Snow Labs +name: t5_abs_qa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_abs_qa_pipeline` is a English model originally trained by tuner007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_abs_qa_pipeline_en_5.4.2_3.0_1724015007089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_abs_qa_pipeline_en_5.4.2_3.0_1724015007089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_abs_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_abs_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_abs_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|988.8 MB| + +## References + +https://huggingface.co/tuner007/t5_abs_qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_en.md new file mode 100644 index 00000000000000..f8e4e551ca453e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_abstract2title T5Transformer from yuewu +author: John Snow Labs +name: t5_abstract2title +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_abstract2title` is a English model originally trained by yuewu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_abstract2title_en_5.4.2_3.0_1723998304621.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_abstract2title_en_5.4.2_3.0_1723998304621.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_abstract2title","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_abstract2title", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_abstract2title| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yuewu/T5_abstract2title \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_pipeline_en.md new file mode 100644 index 00000000000000..2bc6d40bc38d86 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_abstract2title_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_abstract2title_pipeline pipeline T5Transformer from yuewu +author: John Snow Labs +name: t5_abstract2title_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_abstract2title_pipeline` is a English model originally trained by yuewu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_abstract2title_pipeline_en_5.4.2_3.0_1723998350916.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_abstract2title_pipeline_en_5.4.2_3.0_1723998350916.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_abstract2title_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_abstract2title_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_abstract2title_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yuewu/T5_abstract2title + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_en.md new file mode 100644 index 00000000000000..3011df70a32423 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_askscience_lfqa T5Transformer from pszemraj +author: John Snow Labs +name: t5_base_askscience_lfqa +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_askscience_lfqa` is a English model originally trained by pszemraj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_askscience_lfqa_en_5.4.2_3.0_1724020561087.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_askscience_lfqa_en_5.4.2_3.0_1724020561087.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_askscience_lfqa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_askscience_lfqa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_askscience_lfqa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pszemraj/t5-base-askscience-lfqa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_pipeline_en.md new file mode 100644 index 00000000000000..058e63ef338a9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_askscience_lfqa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_askscience_lfqa_pipeline pipeline T5Transformer from pszemraj +author: John Snow Labs +name: t5_base_askscience_lfqa_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_askscience_lfqa_pipeline` is a English model originally trained by pszemraj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_askscience_lfqa_pipeline_en_5.4.2_3.0_1724020612268.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_askscience_lfqa_pipeline_en_5.4.2_3.0_1724020612268.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_askscience_lfqa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_askscience_lfqa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_askscience_lfqa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pszemraj/t5-base-askscience-lfqa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_en.md new file mode 100644 index 00000000000000..b30ef190eff9c0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt0_khanq_eduqg T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt0_khanq_eduqg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt0_khanq_eduqg` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt0_khanq_eduqg_en_5.4.2_3.0_1724001642473.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt0_khanq_eduqg_en_5.4.2_3.0_1724001642473.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt0_khanq_eduqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt0_khanq_eduqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt0_khanq_eduqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt0-khanq-eduqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_pipeline_en.md new file mode 100644 index 00000000000000..85e0ec45c4a4ef --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_bt0_khanq_eduqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt0_khanq_eduqg_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt0_khanq_eduqg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt0_khanq_eduqg_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt0_khanq_eduqg_pipeline_en_5.4.2_3.0_1724001689151.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt0_khanq_eduqg_pipeline_en_5.4.2_3.0_1724001689151.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt0_khanq_eduqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt0_khanq_eduqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt0_khanq_eduqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt0-khanq-eduqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_en.md new file mode 100644 index 00000000000000..6d1a7901e9ff74 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_en_5.4.2_3.0_1723994877122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_en_5.4.2_3.0_1723994877122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_c2_mare_ar1_ex8_half_from_ft_dense_normalization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline_en.md new file mode 100644 index 00000000000000..fa223cde05ecb0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline pipeline T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline_en_5.4.2_3.0_1723995021168.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline_en_5.4.2_3.0_1723995021168.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_c2_mare_ar1_ex8_half_from_ft_dense_normalization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_c2_mare_ar1_ex8_half_from_ft_dense_normalization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_en.md new file mode 100644 index 00000000000000..a43300d9d2bfe8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1723945492415.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1723945492415.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|974.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-1024-finetuned-squad-infilling-seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..33a6914398c1ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1723945552173.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1723945552173.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_1024_finetuned_squad_infilling_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|974.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-1024-finetuned-squad-infilling-seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_en.md new file mode 100644 index 00000000000000..4fb68f38ad3e17 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_256_finetuned_squad_seed_4 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_256_finetuned_squad_seed_4 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_256_finetuned_squad_seed_4` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_256_finetuned_squad_seed_4_en_5.4.2_3.0_1723976774377.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_256_finetuned_squad_seed_4_en_5.4.2_3.0_1723976774377.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_256_finetuned_squad_seed_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_256_finetuned_squad_seed_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_256_finetuned_squad_seed_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|955.6 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-256-finetuned-squad-seed-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline_en.md new file mode 100644 index 00000000000000..c970d8da94a1aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline_en_5.4.2_3.0_1723976841295.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline_en_5.4.2_3.0_1723976841295.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_256_finetuned_squad_seed_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|955.6 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-256-finetuned-squad-seed-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_en.md new file mode 100644 index 00000000000000..ae8eb024e6a7b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_fine_tuned T5Transformer from rob06 +author: John Snow Labs +name: t5_base_fine_tuned +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_fine_tuned` is a English model originally trained by rob06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_fine_tuned_en_5.4.2_3.0_1723942832066.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_fine_tuned_en_5.4.2_3.0_1723942832066.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_fine_tuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_fine_tuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_fine_tuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|939.3 MB| + +## References + +https://huggingface.co/rob06/t5-base-fine-tuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_pipeline_en.md new file mode 100644 index 00000000000000..19a4d73dd88d99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_fine_tuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_fine_tuned_pipeline pipeline T5Transformer from rob06 +author: John Snow Labs +name: t5_base_fine_tuned_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_fine_tuned_pipeline` is a English model originally trained by rob06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_fine_tuned_pipeline_en_5.4.2_3.0_1723942884536.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_fine_tuned_pipeline_en_5.4.2_3.0_1723942884536.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_fine_tuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_fine_tuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_fine_tuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|939.3 MB| + +## References + +https://huggingface.co/rob06/t5-base-fine-tuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_en.md new file mode 100644 index 00000000000000..88c94208f3afd9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_cnn_2 T5Transformer from nizamudma +author: John Snow Labs +name: t5_base_finetuned_cnn_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_cnn_2` is a English model originally trained by nizamudma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_cnn_2_en_5.4.2_3.0_1724004059716.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_cnn_2_en_5.4.2_3.0_1724004059716.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_cnn_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_cnn_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_cnn_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.4 MB| + +## References + +https://huggingface.co/nizamudma/t5-base-finetuned-cnn-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_pipeline_en.md new file mode 100644 index 00000000000000..dcd32c51f9d0a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_cnn_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_cnn_2_pipeline pipeline T5Transformer from nizamudma +author: John Snow Labs +name: t5_base_finetuned_cnn_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_cnn_2_pipeline` is a English model originally trained by nizamudma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_cnn_2_pipeline_en_5.4.2_3.0_1724004062370.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_cnn_2_pipeline_en_5.4.2_3.0_1724004062370.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_cnn_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_cnn_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_cnn_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.4 MB| + +## References + +https://huggingface.co/nizamudma/t5-base-finetuned-cnn-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_en.md new file mode 100644 index 00000000000000..6e1ebbb3551d0e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_math_calculus_differentiate T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_math_calculus_differentiate +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_math_calculus_differentiate` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_math_calculus_differentiate_en_5.4.2_3.0_1724016005144.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_math_calculus_differentiate_en_5.4.2_3.0_1724016005144.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_math_calculus_differentiate","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_math_calculus_differentiate", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_math_calculus_differentiate| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|888.6 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-math-calculus-differentiate \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_pipeline_en.md new file mode 100644 index 00000000000000..5d1a99d40c03ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_math_calculus_differentiate_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_math_calculus_differentiate_pipeline pipeline T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_math_calculus_differentiate_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_math_calculus_differentiate_pipeline` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_math_calculus_differentiate_pipeline_en_5.4.2_3.0_1724016083821.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_math_calculus_differentiate_pipeline_en_5.4.2_3.0_1724016083821.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_math_calculus_differentiate_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_math_calculus_differentiate_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_math_calculus_differentiate_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|888.6 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-math-calculus-differentiate + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_en.md new file mode 100644 index 00000000000000..6e28d970b1a138 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_noun_ellipse T5Transformer from Kartikey95 +author: John Snow Labs +name: t5_base_finetuned_noun_ellipse +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_noun_ellipse` is a English model originally trained by Kartikey95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_noun_ellipse_en_5.4.2_3.0_1724002366677.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_noun_ellipse_en_5.4.2_3.0_1724002366677.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_noun_ellipse","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_noun_ellipse", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_noun_ellipse| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|909.4 MB| + +## References + +https://huggingface.co/Kartikey95/t5-base-finetuned-noun_ellipse \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_pipeline_en.md new file mode 100644 index 00000000000000..ba1e98b84e762e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_noun_ellipse_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_noun_ellipse_pipeline pipeline T5Transformer from Kartikey95 +author: John Snow Labs +name: t5_base_finetuned_noun_ellipse_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_noun_ellipse_pipeline` is a English model originally trained by Kartikey95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_noun_ellipse_pipeline_en_5.4.2_3.0_1724002427404.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_noun_ellipse_pipeline_en_5.4.2_3.0_1724002427404.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_noun_ellipse_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_noun_ellipse_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_noun_ellipse_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|909.4 MB| + +## References + +https://huggingface.co/Kartikey95/t5-base-finetuned-noun_ellipse + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_en.md new file mode 100644 index 00000000000000..0a7c63cc0f940a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_mim T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_mim +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_mim` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_mim_en_5.4.2_3.0_1723959690838.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_mim_en_5.4.2_3.0_1723959690838.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_mim","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_mim", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_mim| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|937.4 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-mim \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline_en.md new file mode 100644 index 00000000000000..7eef4e97780c1c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline pipeline T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline_en_5.4.2_3.0_1723959748121.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline_en_5.4.2_3.0_1723959748121.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_mim_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|937.4 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-mim + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_en.md new file mode 100644 index 00000000000000..c06757fef16d83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_pbb T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_pbb +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_pbb` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pbb_en_5.4.2_3.0_1723948774048.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pbb_en_5.4.2_3.0_1723948774048.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_pbb","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_pbb", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_pbb| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|948.3 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-pbb \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline_en.md new file mode 100644 index 00000000000000..f0675870ad9bda --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline pipeline T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline_en_5.4.2_3.0_1723948826544.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline_en_5.4.2_3.0_1723948826544.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_pbb_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|948.3 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-pbb + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_en.md new file mode 100644 index 00000000000000..fffa02c80a4880 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_pua T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_pua +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_pua` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pua_en_5.4.2_3.0_1723995727738.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pua_en_5.4.2_3.0_1723995727738.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_pua","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_spanish_tonga_tonga_islands_pua", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_pua| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|949.7 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-pua \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline_en.md new file mode 100644 index 00000000000000..a26656983d5392 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline pipeline T5Transformer from mekjr1 +author: John Snow Labs +name: t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline` is a English model originally trained by mekjr1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline_en_5.4.2_3.0_1723995777788.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline_en_5.4.2_3.0_1723995777788.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spanish_tonga_tonga_islands_pua_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|949.7 MB| + +## References + +https://huggingface.co/mekjr1/t5-base-finetuned-es-to-pua + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_en.md new file mode 100644 index 00000000000000..355a94c91b4031 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_tab_fact T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_tab_fact +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_tab_fact` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tab_fact_en_5.4.2_3.0_1723959155575.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tab_fact_en_5.4.2_3.0_1723959155575.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_tab_fact","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_tab_fact", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_tab_fact| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|952.0 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-tab_fact \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_pipeline_en.md new file mode 100644 index 00000000000000..814ab51cb2bbfb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_finetuned_tab_fact_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_tab_fact_pipeline pipeline T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_tab_fact_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_tab_fact_pipeline` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tab_fact_pipeline_en_5.4.2_3.0_1723959220860.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tab_fact_pipeline_en_5.4.2_3.0_1723959220860.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_tab_fact_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_tab_fact_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_tab_fact_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|952.0 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-tab_fact + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_en.md new file mode 100644 index 00000000000000..04e72b7807fbba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_mediasum_seed17 T5Transformer from PSW +author: John Snow Labs +name: t5_base_mediasum_seed17 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_mediasum_seed17` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed17_en_5.4.2_3.0_1724001648954.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed17_en_5.4.2_3.0_1724001648954.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_mediasum_seed17","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_mediasum_seed17", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_mediasum_seed17| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-mediasum-seed17 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_pipeline_en.md new file mode 100644 index 00000000000000..0998d8f102e1c6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_mediasum_seed17_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_mediasum_seed17_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_mediasum_seed17_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_mediasum_seed17_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed17_pipeline_en_5.4.2_3.0_1724001696495.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_mediasum_seed17_pipeline_en_5.4.2_3.0_1724001696495.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_mediasum_seed17_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_mediasum_seed17_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_mediasum_seed17_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-mediasum-seed17 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_ja.md new file mode 100644 index 00000000000000..9ca31a3da2e577 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_ja.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Japanese t5_base_medium T5Transformer from retrieva-jp +author: John Snow Labs +name: t5_base_medium +date: 2024-08-18 +tags: [ja, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_medium` is a Japanese model originally trained by retrieva-jp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_medium_ja_5.4.2_3.0_1724023647212.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_medium_ja_5.4.2_3.0_1724023647212.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_medium","ja") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_medium", "ja") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_medium| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|1.0 GB| + +## References + +https://huggingface.co/retrieva-jp/t5-base-medium \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_pipeline_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_pipeline_ja.md new file mode 100644 index 00000000000000..c0a1c5dac7ca80 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_medium_pipeline_ja.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Japanese t5_base_medium_pipeline pipeline T5Transformer from retrieva-jp +author: John Snow Labs +name: t5_base_medium_pipeline +date: 2024-08-18 +tags: [ja, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_medium_pipeline` is a Japanese model originally trained by retrieva-jp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_medium_pipeline_ja_5.4.2_3.0_1724023691416.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_medium_pipeline_ja_5.4.2_3.0_1724023691416.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_medium_pipeline", lang = "ja") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_medium_pipeline", lang = "ja") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_medium_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ja| +|Size:|1.0 GB| + +## References + +https://huggingface.co/retrieva-jp/t5-base-medium + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_en.md new file mode 100644 index 00000000000000..7d5180ea63f4b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_newsela_biendata_with_domain_adaptation T5Transformer from bogdancazan +author: John Snow Labs +name: t5_base_newsela_biendata_with_domain_adaptation +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_newsela_biendata_with_domain_adaptation` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_newsela_biendata_with_domain_adaptation_en_5.4.2_3.0_1723947123411.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_newsela_biendata_with_domain_adaptation_en_5.4.2_3.0_1723947123411.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_newsela_biendata_with_domain_adaptation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_newsela_biendata_with_domain_adaptation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_newsela_biendata_with_domain_adaptation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bogdancazan/t5-base-newsela-biendata-with-domain-adaptation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_pipeline_en.md new file mode 100644 index 00000000000000..b0d2b57c32e45b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_newsela_biendata_with_domain_adaptation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_newsela_biendata_with_domain_adaptation_pipeline pipeline T5Transformer from bogdancazan +author: John Snow Labs +name: t5_base_newsela_biendata_with_domain_adaptation_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_newsela_biendata_with_domain_adaptation_pipeline` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_newsela_biendata_with_domain_adaptation_pipeline_en_5.4.2_3.0_1723947174920.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_newsela_biendata_with_domain_adaptation_pipeline_en_5.4.2_3.0_1723947174920.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_newsela_biendata_with_domain_adaptation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_newsela_biendata_with_domain_adaptation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_newsela_biendata_with_domain_adaptation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bogdancazan/t5-base-newsela-biendata-with-domain-adaptation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_en.md new file mode 100644 index 00000000000000..7837bad39a73d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_parasci T5Transformer from HelloRusk +author: John Snow Labs +name: t5_base_parasci +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_parasci` is a English model originally trained by HelloRusk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_parasci_en_5.4.2_3.0_1724023015968.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_parasci_en_5.4.2_3.0_1724023015968.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_parasci","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_parasci", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_parasci| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|997.8 MB| + +## References + +https://huggingface.co/HelloRusk/t5-base-parasci \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_pipeline_en.md new file mode 100644 index 00000000000000..a757ba56c92f63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_parasci_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_parasci_pipeline pipeline T5Transformer from HelloRusk +author: John Snow Labs +name: t5_base_parasci_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_parasci_pipeline` is a English model originally trained by HelloRusk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_parasci_pipeline_en_5.4.2_3.0_1724023065670.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_parasci_pipeline_en_5.4.2_3.0_1724023065670.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_parasci_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_parasci_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_parasci_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|997.8 MB| + +## References + +https://huggingface.co/HelloRusk/t5-base-parasci + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_en.md new file mode 100644 index 00000000000000..1d23f7fc82908f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_prompter_multiarith_300_repeated_ep1 T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_prompter_multiarith_300_repeated_ep1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_prompter_multiarith_300_repeated_ep1` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_prompter_multiarith_300_repeated_ep1_en_5.4.2_3.0_1723976116544.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_prompter_multiarith_300_repeated_ep1_en_5.4.2_3.0_1723976116544.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_prompter_multiarith_300_repeated_ep1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_prompter_multiarith_300_repeated_ep1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_prompter_multiarith_300_repeated_ep1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|918.8 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-prompter-multiarith_300-repeated-ep1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_pipeline_en.md new file mode 100644 index 00000000000000..42201f2ffe0b57 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_prompter_multiarith_300_repeated_ep1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_prompter_multiarith_300_repeated_ep1_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_prompter_multiarith_300_repeated_ep1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_prompter_multiarith_300_repeated_ep1_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_prompter_multiarith_300_repeated_ep1_pipeline_en_5.4.2_3.0_1723976177383.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_prompter_multiarith_300_repeated_ep1_pipeline_en_5.4.2_3.0_1723976177383.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_prompter_multiarith_300_repeated_ep1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_prompter_multiarith_300_repeated_ep1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_prompter_multiarith_300_repeated_ep1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|918.8 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-prompter-multiarith_300-repeated-ep1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_en.md new file mode 100644 index 00000000000000..82f4fba5c881a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_amazon_beauty T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_amazon_beauty +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_amazon_beauty` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_amazon_beauty_en_5.4.2_3.0_1723982317492.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_amazon_beauty_en_5.4.2_3.0_1723982317492.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_amazon_beauty","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_amazon_beauty", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_amazon_beauty| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|997.1 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-amazon-beauty \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_pipeline_en.md new file mode 100644 index 00000000000000..0f12e2e4fd82a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_amazon_beauty_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_amazon_beauty_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_amazon_beauty_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_amazon_beauty_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_amazon_beauty_pipeline_en_5.4.2_3.0_1723982371271.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_amazon_beauty_pipeline_en_5.4.2_3.0_1723982371271.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_amazon_beauty_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_amazon_beauty_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_amazon_beauty_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|997.1 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-amazon-beauty + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_en.md new file mode 100644 index 00000000000000..f0d6cc0081b175 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_electronics T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_electronics +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_electronics` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_electronics_en_5.4.2_3.0_1723979853553.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_electronics_en_5.4.2_3.0_1723979853553.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_electronics","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_electronics", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_electronics| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.8 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-electronics \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_pipeline_en.md new file mode 100644 index 00000000000000..b401846cdb1e13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_electronics_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_electronics_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_electronics_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_electronics_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_electronics_pipeline_en_5.4.2_3.0_1723979906768.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_electronics_pipeline_en_5.4.2_3.0_1723979906768.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_bm25_electronics_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_bm25_electronics_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_electronics_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.8 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-electronics + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_en.md new file mode 100644 index 00000000000000..4edd005b5421e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_toys T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_toys +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_toys` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_toys_en_5.4.2_3.0_1723954165878.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_toys_en_5.4.2_3.0_1723954165878.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_toys","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_bm25_toys", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_toys| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|970.5 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-toys \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_pipeline_en.md new file mode 100644 index 00000000000000..80cef462e0d469 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_rlhf_bm25_toys_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_bm25_toys_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_bm25_toys_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_bm25_toys_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_toys_pipeline_en_5.4.2_3.0_1723954226326.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_bm25_toys_pipeline_en_5.4.2_3.0_1723954226326.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_bm25_toys_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_bm25_toys_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_bm25_toys_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|970.5 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-bm25-toys + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_en.md new file mode 100644 index 00000000000000..b3357c06f0f7f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sft_pet T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_pet +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_pet` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_pet_en_5.4.2_3.0_1724003396137.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_pet_en_5.4.2_3.0_1724003396137.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sft_pet","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sft_pet", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_pet| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|986.2 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-pet \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_pipeline_en.md new file mode 100644 index 00000000000000..65ebf318054019 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sft_pet_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sft_pet_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_pet_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_pet_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_pet_pipeline_en_5.4.2_3.0_1724003448826.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_pet_pipeline_en_5.4.2_3.0_1724003448826.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sft_pet_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sft_pet_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_pet_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|986.2 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-pet + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_ja.md new file mode 100644 index 00000000000000..e2c0770240a174 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_ja.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Japanese t5_base_short T5Transformer from retrieva-jp +author: John Snow Labs +name: t5_base_short +date: 2024-08-18 +tags: [ja, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_short` is a Japanese model originally trained by retrieva-jp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_short_ja_5.4.2_3.0_1724020707803.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_short_ja_5.4.2_3.0_1724020707803.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_short","ja") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_short", "ja") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_short| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|1.0 GB| + +## References + +https://huggingface.co/retrieva-jp/t5-base-short \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_pipeline_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_pipeline_ja.md new file mode 100644 index 00000000000000..7d7419882c5441 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_short_pipeline_ja.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Japanese t5_base_short_pipeline pipeline T5Transformer from retrieva-jp +author: John Snow Labs +name: t5_base_short_pipeline +date: 2024-08-18 +tags: [ja, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_short_pipeline` is a Japanese model originally trained by retrieva-jp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_short_pipeline_ja_5.4.2_3.0_1724020755251.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_short_pipeline_ja_5.4.2_3.0_1724020755251.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_short_pipeline", lang = "ja") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_short_pipeline", lang = "ja") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_short_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ja| +|Size:|1.0 GB| + +## References + +https://huggingface.co/retrieva-jp/t5-base-short + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_en.md new file mode 100644 index 00000000000000..2678283ca151dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_squadshifts_nyt_qg T5Transformer from research-backup +author: John Snow Labs +name: t5_base_squadshifts_nyt_qg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squadshifts_nyt_qg` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_nyt_qg_en_5.4.2_3.0_1723979690380.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_nyt_qg_en_5.4.2_3.0_1723979690380.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_squadshifts_nyt_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_squadshifts_nyt_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squadshifts_nyt_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-squadshifts-nyt-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_pipeline_en.md new file mode 100644 index 00000000000000..662dced02e8a92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_nyt_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_squadshifts_nyt_qg_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_base_squadshifts_nyt_qg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squadshifts_nyt_qg_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_nyt_qg_pipeline_en_5.4.2_3.0_1723979738679.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_nyt_qg_pipeline_en_5.4.2_3.0_1723979738679.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_squadshifts_nyt_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_squadshifts_nyt_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squadshifts_nyt_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-squadshifts-nyt-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_en.md new file mode 100644 index 00000000000000..338339cc5172f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_squadshifts_vanilla_amazon_qg T5Transformer from research-backup +author: John Snow Labs +name: t5_base_squadshifts_vanilla_amazon_qg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squadshifts_vanilla_amazon_qg` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_vanilla_amazon_qg_en_5.4.2_3.0_1723995689565.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_vanilla_amazon_qg_en_5.4.2_3.0_1723995689565.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_squadshifts_vanilla_amazon_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_squadshifts_vanilla_amazon_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squadshifts_vanilla_amazon_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-squadshifts-vanilla-amazon-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_pipeline_en.md new file mode 100644 index 00000000000000..ac54b2cf45ac1a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_squadshifts_vanilla_amazon_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_squadshifts_vanilla_amazon_qg_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_base_squadshifts_vanilla_amazon_qg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squadshifts_vanilla_amazon_qg_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_vanilla_amazon_qg_pipeline_en_5.4.2_3.0_1723995734972.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squadshifts_vanilla_amazon_qg_pipeline_en_5.4.2_3.0_1723995734972.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_squadshifts_vanilla_amazon_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_squadshifts_vanilla_amazon_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squadshifts_vanilla_amazon_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-squadshifts-vanilla-amazon-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_en.md new file mode 100644 index 00000000000000..2ce3eb1cdbd7b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sushantgautam T5Transformer from SushantGautam +author: John Snow Labs +name: t5_base_sushantgautam +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sushantgautam` is a English model originally trained by SushantGautam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sushantgautam_en_5.4.2_3.0_1723942158780.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sushantgautam_en_5.4.2_3.0_1723942158780.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sushantgautam","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sushantgautam", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sushantgautam| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SushantGautam/t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_pipeline_en.md new file mode 100644 index 00000000000000..33c4c13318f35f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_sushantgautam_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sushantgautam_pipeline pipeline T5Transformer from SushantGautam +author: John Snow Labs +name: t5_base_sushantgautam_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sushantgautam_pipeline` is a English model originally trained by SushantGautam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sushantgautam_pipeline_en_5.4.2_3.0_1723942210453.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sushantgautam_pipeline_en_5.4.2_3.0_1723942210453.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sushantgautam_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sushantgautam_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sushantgautam_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SushantGautam/t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_en.md new file mode 100644 index 00000000000000..42705561017050 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_10front_1body_0rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_10front_1body_0rear +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_10front_1body_0rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_10front_1body_0rear_en_5.4.2_3.0_1723946376636.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_10front_1body_0rear_en_5.4.2_3.0_1723946376636.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_10front_1body_0rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_10front_1body_0rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_10front_1body_0rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-10front-1body-0rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_pipeline_en.md new file mode 100644 index 00000000000000..058d65e8cd8afd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_10front_1body_0rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_10front_1body_0rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_10front_1body_0rear_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_10front_1body_0rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_10front_1body_0rear_pipeline_en_5.4.2_3.0_1723946429733.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_10front_1body_0rear_pipeline_en_5.4.2_3.0_1723946429733.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_10front_1body_0rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_10front_1body_0rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_10front_1body_0rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-10front-1body-0rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_en.md new file mode 100644 index 00000000000000..a492c719ebd6c1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_9front_1body_0rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_9front_1body_0rear +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_9front_1body_0rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_0rear_en_5.4.2_3.0_1723973653651.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_0rear_en_5.4.2_3.0_1723973653651.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_9front_1body_0rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_9front_1body_0rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_9front_1body_0rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-9front-1body-0rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_pipeline_en.md new file mode 100644 index 00000000000000..727f1fb5ad91dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_tedxjp_9front_1body_0rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_9front_1body_0rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_9front_1body_0rear_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_9front_1body_0rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_0rear_pipeline_en_5.4.2_3.0_1723973701704.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_0rear_pipeline_en_5.4.2_3.0_1723973701704.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_9front_1body_0rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_9front_1body_0rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_9front_1body_0rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-9front-1body-0rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_en.md new file mode 100644 index 00000000000000..78eb82f7376eaa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_title_v4_ashishkgpian T5Transformer from ashishkgpian +author: John Snow Labs +name: t5_base_title_v4_ashishkgpian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_title_v4_ashishkgpian` is a English model originally trained by ashishkgpian. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_title_v4_ashishkgpian_en_5.4.2_3.0_1723942344879.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_title_v4_ashishkgpian_en_5.4.2_3.0_1723942344879.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_title_v4_ashishkgpian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_title_v4_ashishkgpian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_title_v4_ashishkgpian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ashishkgpian/T5_base_title_v4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_pipeline_en.md new file mode 100644 index 00000000000000..c659d6759f06f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_title_v4_ashishkgpian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_title_v4_ashishkgpian_pipeline pipeline T5Transformer from ashishkgpian +author: John Snow Labs +name: t5_base_title_v4_ashishkgpian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_title_v4_ashishkgpian_pipeline` is a English model originally trained by ashishkgpian. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_title_v4_ashishkgpian_pipeline_en_5.4.2_3.0_1723942396668.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_title_v4_ashishkgpian_pipeline_en_5.4.2_3.0_1723942396668.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_title_v4_ashishkgpian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_title_v4_ashishkgpian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_title_v4_ashishkgpian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ashishkgpian/T5_base_title_v4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_en.md new file mode 100644 index 00000000000000..e6338fc61f4c1e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_ukrainian_tonga_tonga_islands_us_english T5Transformer from EnglishVoice +author: John Snow Labs +name: t5_base_ukrainian_tonga_tonga_islands_us_english +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_ukrainian_tonga_tonga_islands_us_english` is a English model originally trained by EnglishVoice. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_ukrainian_tonga_tonga_islands_us_english_en_5.4.2_3.0_1724012737740.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_ukrainian_tonga_tonga_islands_us_english_en_5.4.2_3.0_1724012737740.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_ukrainian_tonga_tonga_islands_us_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_ukrainian_tonga_tonga_islands_us_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_ukrainian_tonga_tonga_islands_us_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/EnglishVoice/t5-base-uk-to-us-english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline_en.md new file mode 100644 index 00000000000000..d8265f498b4c26 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline pipeline T5Transformer from EnglishVoice +author: John Snow Labs +name: t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline` is a English model originally trained by EnglishVoice. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline_en_5.4.2_3.0_1724012782864.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline_en_5.4.2_3.0_1724012782864.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_ukrainian_tonga_tonga_islands_us_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/EnglishVoice/t5-base-uk-to-us-english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_en.md new file mode 100644 index 00000000000000..7fa00599dfbf48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_us_tonga_tonga_islands_ukrainian_english T5Transformer from EnglishVoice +author: John Snow Labs +name: t5_base_us_tonga_tonga_islands_ukrainian_english +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_us_tonga_tonga_islands_ukrainian_english` is a English model originally trained by EnglishVoice. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_us_tonga_tonga_islands_ukrainian_english_en_5.4.2_3.0_1724015321640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_us_tonga_tonga_islands_ukrainian_english_en_5.4.2_3.0_1724015321640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_us_tonga_tonga_islands_ukrainian_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_us_tonga_tonga_islands_ukrainian_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_us_tonga_tonga_islands_ukrainian_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/EnglishVoice/t5-base-us-to-uk-english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline_en.md new file mode 100644 index 00000000000000..6af186200a17b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline pipeline T5Transformer from EnglishVoice +author: John Snow Labs +name: t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline` is a English model originally trained by EnglishVoice. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline_en_5.4.2_3.0_1724015370683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline_en_5.4.2_3.0_1724015370683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_us_tonga_tonga_islands_ukrainian_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/EnglishVoice/t5-base-us-to-uk-english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_en.md new file mode 100644 index 00000000000000..a11aa7b8d26f0f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_with_title_multi_german_wiki_news_airklizz T5Transformer from airKlizz +author: John Snow Labs +name: t5_base_with_title_multi_german_wiki_news_airklizz +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_with_title_multi_german_wiki_news_airklizz` is a English model originally trained by airKlizz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_with_title_multi_german_wiki_news_airklizz_en_5.4.2_3.0_1724024238640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_with_title_multi_german_wiki_news_airklizz_en_5.4.2_3.0_1724024238640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_with_title_multi_german_wiki_news_airklizz","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_with_title_multi_german_wiki_news_airklizz", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_with_title_multi_german_wiki_news_airklizz| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|986.1 MB| + +## References + +https://huggingface.co/airKlizz/t5-base-with-title-multi-de-wiki-news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_pipeline_en.md new file mode 100644 index 00000000000000..eaa49b86035b20 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_base_with_title_multi_german_wiki_news_airklizz_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_with_title_multi_german_wiki_news_airklizz_pipeline pipeline T5Transformer from airKlizz +author: John Snow Labs +name: t5_base_with_title_multi_german_wiki_news_airklizz_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_with_title_multi_german_wiki_news_airklizz_pipeline` is a English model originally trained by airKlizz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_with_title_multi_german_wiki_news_airklizz_pipeline_en_5.4.2_3.0_1724024291955.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_with_title_multi_german_wiki_news_airklizz_pipeline_en_5.4.2_3.0_1724024291955.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_with_title_multi_german_wiki_news_airklizz_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_with_title_multi_german_wiki_news_airklizz_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_with_title_multi_german_wiki_news_airklizz_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|986.1 MB| + +## References + +https://huggingface.co/airKlizz/t5-base-with-title-multi-de-wiki-news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_en.md new file mode 100644 index 00000000000000..18643c0b49d044 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_coktails_recipe_small T5Transformer from erwanlc +author: John Snow Labs +name: t5_coktails_recipe_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_coktails_recipe_small` is a English model originally trained by erwanlc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_coktails_recipe_small_en_5.4.2_3.0_1723940499715.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_coktails_recipe_small_en_5.4.2_3.0_1723940499715.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_coktails_recipe_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_coktails_recipe_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_coktails_recipe_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|321.8 MB| + +## References + +https://huggingface.co/erwanlc/t5-coktails_recipe-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_pipeline_en.md new file mode 100644 index 00000000000000..c78efb03ac308f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_coktails_recipe_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_coktails_recipe_small_pipeline pipeline T5Transformer from erwanlc +author: John Snow Labs +name: t5_coktails_recipe_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_coktails_recipe_small_pipeline` is a English model originally trained by erwanlc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_coktails_recipe_small_pipeline_en_5.4.2_3.0_1723940518126.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_coktails_recipe_small_pipeline_en_5.4.2_3.0_1723940518126.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_coktails_recipe_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_coktails_recipe_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_coktails_recipe_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|321.8 MB| + +## References + +https://huggingface.co/erwanlc/t5-coktails_recipe-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_en.md new file mode 100644 index 00000000000000..8fb50a8925c81b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_detox_batalovme T5Transformer from batalovme +author: John Snow Labs +name: t5_detox_batalovme +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_detox_batalovme` is a English model originally trained by batalovme. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_detox_batalovme_en_5.4.2_3.0_1724005425733.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_detox_batalovme_en_5.4.2_3.0_1724005425733.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_detox_batalovme","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_detox_batalovme", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_detox_batalovme| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/batalovme/t5-detox \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_pipeline_en.md new file mode 100644 index 00000000000000..f779820d82b230 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_detox_batalovme_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_detox_batalovme_pipeline pipeline T5Transformer from batalovme +author: John Snow Labs +name: t5_detox_batalovme_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_detox_batalovme_pipeline` is a English model originally trained by batalovme. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_detox_batalovme_pipeline_en_5.4.2_3.0_1724005472870.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_detox_batalovme_pipeline_en_5.4.2_3.0_1724005472870.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_detox_batalovme_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_detox_batalovme_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_detox_batalovme_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/batalovme/t5-detox + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_en.md new file mode 100644 index 00000000000000..e8e761ab391cb9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20 T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_en_5.4.2_3.0_1723952669084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_en_5.4.2_3.0_1723952669084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|992.0 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-1PLUSalpha0-9-e20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline_en.md new file mode 100644 index 00000000000000..ec17a3fca9db7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline pipeline T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline_en_5.4.2_3.0_1723952726428.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline_en_5.4.2_3.0_1723952726428.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|992.0 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-1PLUSalpha0-9-e20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_en.md new file mode 100644 index 00000000000000..e418bcf15c7d8b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30 T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_en_5.4.2_3.0_1724007418994.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_en_5.4.2_3.0_1724007418994.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|993.2 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-1PLUSalpha0-9-e30 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline_en.md new file mode 100644 index 00000000000000..a19021ea71e7f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline pipeline T5Transformer from harish +author: John Snow Labs +name: t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline_en_5.4.2_3.0_1724007469749.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline_en_5.4.2_3.0_1724007469749.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_e2e_10epochs_lr1e4_alpha0_1plusalpha0_9_e30_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|993.2 MB| + +## References + +https://huggingface.co/harish/t5-e2e-10epochs-lr1e4-alpha0-1PLUSalpha0-9-e30 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_de.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_de.md new file mode 100644 index 00000000000000..bcf6799b9f3b3b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_de.md @@ -0,0 +1,86 @@ +--- +layout: model +title: German t5_efficient_gc4_all_german_small_el32_o5011 T5Transformer from o5011 +author: John Snow Labs +name: t5_efficient_gc4_all_german_small_el32_o5011 +date: 2024-08-18 +tags: [de, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_gc4_all_german_small_el32_o5011` is a German model originally trained by o5011. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_all_german_small_el32_o5011_de_5.4.2_3.0_1723951314245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_all_german_small_el32_o5011_de_5.4.2_3.0_1723951314245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_gc4_all_german_small_el32_o5011","de") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_gc4_all_german_small_el32_o5011", "de") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_gc4_all_german_small_el32_o5011| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|de| +|Size:|337.1 MB| + +## References + +https://huggingface.co/o5011/t5-efficient-gc4-all-german-small-el32 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_pipeline_de.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_pipeline_de.md new file mode 100644 index 00000000000000..c7a28524a98455 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_gc4_all_german_small_el32_o5011_pipeline_de.md @@ -0,0 +1,69 @@ +--- +layout: model +title: German t5_efficient_gc4_all_german_small_el32_o5011_pipeline pipeline T5Transformer from o5011 +author: John Snow Labs +name: t5_efficient_gc4_all_german_small_el32_o5011_pipeline +date: 2024-08-18 +tags: [de, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_gc4_all_german_small_el32_o5011_pipeline` is a German model originally trained by o5011. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_all_german_small_el32_o5011_pipeline_de_5.4.2_3.0_1723951429236.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_all_german_small_el32_o5011_pipeline_de_5.4.2_3.0_1723951429236.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_gc4_all_german_small_el32_o5011_pipeline", lang = "de") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_gc4_all_german_small_el32_o5011_pipeline", lang = "de") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_gc4_all_german_small_el32_o5011_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|de| +|Size:|337.1 MB| + +## References + +https://huggingface.co/o5011/t5-efficient-gc4-all-german-small-el32 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_en.md new file mode 100644 index 00000000000000..638253f4a7495e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_en.md @@ -0,0 +1,94 @@ +--- +layout: model +title: English T5ForConditionalGeneration Small Cased model (from google) +author: John Snow Labs +name: t5_efficient_small_kv256 +date: 2024-08-18 +tags: [en, open_source, t5, onnx] +task: Text Generation +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `t5-efficient-small-kv256` is a English model originally trained by `google`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_small_kv256_en_5.4.2_3.0_1724008921898.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_small_kv256_en_5.4.2_3.0_1724008921898.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_efficient_small_kv256","en") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_small_kv256","en") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_small_kv256| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|287.1 MB| + +## References + +References + +- https://huggingface.co/google/t5-efficient-small-kv256 +- https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html +- https://arxiv.org/abs/2109.10686 +- https://arxiv.org/abs/2109.10686 +- https://github.com/google-research/google-research/issues/986#issuecomment-1035051145 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_pipeline_en.md new file mode 100644 index 00000000000000..e1d95bad26dd8d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_kv256_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_small_kv256_pipeline pipeline T5Transformer from google +author: John Snow Labs +name: t5_efficient_small_kv256_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_small_kv256_pipeline` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_small_kv256_pipeline_en_5.4.2_3.0_1724009015620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_small_kv256_pipeline_en_5.4.2_3.0_1724009015620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_small_kv256_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_small_kv256_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_small_kv256_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|287.1 MB| + +## References + +https://huggingface.co/google/t5-efficient-small-kv256 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_en.md new file mode 100644 index 00000000000000..41b8d96b64664d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_small_nl16_samsum_exp1 T5Transformer from Gozdi +author: John Snow Labs +name: t5_efficient_small_nl16_samsum_exp1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_small_nl16_samsum_exp1` is a English model originally trained by Gozdi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_small_nl16_samsum_exp1_en_5.4.2_3.0_1723939327499.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_small_nl16_samsum_exp1_en_5.4.2_3.0_1723939327499.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_small_nl16_samsum_exp1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_small_nl16_samsum_exp1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_small_nl16_samsum_exp1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|619.7 MB| + +## References + +https://huggingface.co/Gozdi/t5-efficient-small-nl16-samsum-exp1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_pipeline_en.md new file mode 100644 index 00000000000000..7d58a4ca14e201 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_efficient_small_nl16_samsum_exp1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_small_nl16_samsum_exp1_pipeline pipeline T5Transformer from Gozdi +author: John Snow Labs +name: t5_efficient_small_nl16_samsum_exp1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_small_nl16_samsum_exp1_pipeline` is a English model originally trained by Gozdi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_small_nl16_samsum_exp1_pipeline_en_5.4.2_3.0_1723939358220.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_small_nl16_samsum_exp1_pipeline_en_5.4.2_3.0_1723939358220.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_small_nl16_samsum_exp1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_small_nl16_samsum_exp1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_small_nl16_samsum_exp1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|619.7 MB| + +## References + +https://huggingface.co/Gozdi/t5-efficient-small-nl16-samsum-exp1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_en.md new file mode 100644 index 00000000000000..537034ef2ded1f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_fine_tune_save_example_shashivish T5Transformer from ShashiVish +author: John Snow Labs +name: t5_fine_tune_save_example_shashivish +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_fine_tune_save_example_shashivish` is a English model originally trained by ShashiVish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_fine_tune_save_example_shashivish_en_5.4.2_3.0_1723975909310.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_fine_tune_save_example_shashivish_en_5.4.2_3.0_1723975909310.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_fine_tune_save_example_shashivish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_fine_tune_save_example_shashivish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_fine_tune_save_example_shashivish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.7 MB| + +## References + +https://huggingface.co/ShashiVish/t5-fine-tune-save-example \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_pipeline_en.md new file mode 100644 index 00000000000000..9712b71d400225 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tune_save_example_shashivish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_fine_tune_save_example_shashivish_pipeline pipeline T5Transformer from ShashiVish +author: John Snow Labs +name: t5_fine_tune_save_example_shashivish_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_fine_tune_save_example_shashivish_pipeline` is a English model originally trained by ShashiVish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_fine_tune_save_example_shashivish_pipeline_en_5.4.2_3.0_1723975927844.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_fine_tune_save_example_shashivish_pipeline_en_5.4.2_3.0_1723975927844.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_fine_tune_save_example_shashivish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_fine_tune_save_example_shashivish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_fine_tune_save_example_shashivish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.7 MB| + +## References + +https://huggingface.co/ShashiVish/t5-fine-tune-save-example + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_en.md new file mode 100644 index 00000000000000..d1c1e3064f1aaf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_fine_tuned_300step_f1 T5Transformer from Yeobin +author: John Snow Labs +name: t5_fine_tuned_300step_f1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_fine_tuned_300step_f1` is a English model originally trained by Yeobin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_fine_tuned_300step_f1_en_5.4.2_3.0_1723965955952.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_fine_tuned_300step_f1_en_5.4.2_3.0_1723965955952.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_fine_tuned_300step_f1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_fine_tuned_300step_f1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_fine_tuned_300step_f1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|540.5 MB| + +## References + +https://huggingface.co/Yeobin/t5-fine-tuned-300step-f1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_pipeline_en.md new file mode 100644 index 00000000000000..b8f3b8d26dea05 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_fine_tuned_300step_f1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_fine_tuned_300step_f1_pipeline pipeline T5Transformer from Yeobin +author: John Snow Labs +name: t5_fine_tuned_300step_f1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_fine_tuned_300step_f1_pipeline` is a English model originally trained by Yeobin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_fine_tuned_300step_f1_pipeline_en_5.4.2_3.0_1723966135302.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_fine_tuned_300step_f1_pipeline_en_5.4.2_3.0_1723966135302.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_fine_tuned_300step_f1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_fine_tuned_300step_f1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_fine_tuned_300step_f1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|540.5 MB| + +## References + +https://huggingface.co/Yeobin/t5-fine-tuned-300step-f1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_en.md new file mode 100644 index 00000000000000..6eec9f1dbc0ecc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_finetuned T5Transformer from elinaparajuli +author: John Snow Labs +name: t5_finetuned_finetuned +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_finetuned` is a English model originally trained by elinaparajuli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_finetuned_en_5.4.2_3.0_1723984242711.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_finetuned_en_5.4.2_3.0_1723984242711.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|296.2 MB| + +## References + +https://huggingface.co/elinaparajuli/T5_Finetuned-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..b2e4cfe0a27fa5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_finetuned_pipeline pipeline T5Transformer from elinaparajuli +author: John Snow Labs +name: t5_finetuned_finetuned_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_finetuned_pipeline` is a English model originally trained by elinaparajuli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_finetuned_pipeline_en_5.4.2_3.0_1723984269230.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_finetuned_pipeline_en_5.4.2_3.0_1723984269230.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|296.2 MB| + +## References + +https://huggingface.co/elinaparajuli/T5_Finetuned-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_en.md new file mode 100644 index 00000000000000..c97e6c9a962832 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_genboolq T5Transformer from nc33 +author: John Snow Labs +name: t5_finetuned_genboolq +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_genboolq` is a English model originally trained by nc33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_genboolq_en_5.4.2_3.0_1724005254617.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_genboolq_en_5.4.2_3.0_1724005254617.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_genboolq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_genboolq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_genboolq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/nc33/t5_finetuned_genboolq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_pipeline_en.md new file mode 100644 index 00000000000000..3a5d07a12612e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_finetuned_genboolq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_genboolq_pipeline pipeline T5Transformer from nc33 +author: John Snow Labs +name: t5_finetuned_genboolq_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_genboolq_pipeline` is a English model originally trained by nc33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_genboolq_pipeline_en_5.4.2_3.0_1724005305583.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_genboolq_pipeline_en_5.4.2_3.0_1724005305583.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_genboolq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_genboolq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_genboolq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|998.6 MB| + +## References + +https://huggingface.co/nc33/t5_finetuned_genboolq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_en.md new file mode 100644 index 00000000000000..4a117cc0fe6189 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_flan_small_english_filipino512012 T5Transformer from Cheeseka +author: John Snow Labs +name: t5_flan_small_english_filipino512012 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_small_english_filipino512012` is a English model originally trained by Cheeseka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_small_english_filipino512012_en_5.4.2_3.0_1723996173052.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_small_english_filipino512012_en_5.4.2_3.0_1723996173052.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_flan_small_english_filipino512012","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_flan_small_english_filipino512012", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_small_english_filipino512012| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|346.1 MB| + +## References + +https://huggingface.co/Cheeseka/t5_flan_small_english_filipino512012 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_pipeline_en.md new file mode 100644 index 00000000000000..17fcda57f5ddb1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_flan_small_english_filipino512012_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_flan_small_english_filipino512012_pipeline pipeline T5Transformer from Cheeseka +author: John Snow Labs +name: t5_flan_small_english_filipino512012_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_small_english_filipino512012_pipeline` is a English model originally trained by Cheeseka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_small_english_filipino512012_pipeline_en_5.4.2_3.0_1723996192797.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_small_english_filipino512012_pipeline_en_5.4.2_3.0_1723996192797.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_flan_small_english_filipino512012_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_flan_small_english_filipino512012_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_small_english_filipino512012_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|346.1 MB| + +## References + +https://huggingface.co/Cheeseka/t5_flan_small_english_filipino512012 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_en.md new file mode 100644 index 00000000000000..d9eb68264cb05e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_for_translation_zamachi T5Transformer from Zamachi +author: John Snow Labs +name: t5_for_translation_zamachi +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_for_translation_zamachi` is a English model originally trained by Zamachi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_for_translation_zamachi_en_5.4.2_3.0_1723963987000.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_for_translation_zamachi_en_5.4.2_3.0_1723963987000.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_for_translation_zamachi","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_for_translation_zamachi", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_for_translation_zamachi| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/Zamachi/t5-for-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_pipeline_en.md new file mode 100644 index 00000000000000..7ec1a6a5d2d1d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_for_translation_zamachi_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_for_translation_zamachi_pipeline pipeline T5Transformer from Zamachi +author: John Snow Labs +name: t5_for_translation_zamachi_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_for_translation_zamachi_pipeline` is a English model originally trained by Zamachi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_for_translation_zamachi_pipeline_en_5.4.2_3.0_1723964004092.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_for_translation_zamachi_pipeline_en_5.4.2_3.0_1723964004092.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_for_translation_zamachi_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_for_translation_zamachi_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_for_translation_zamachi_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/Zamachi/t5-for-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_en.md new file mode 100644 index 00000000000000..cc71df06c7afa0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_grammar_model18000 T5Transformer from kalobiralo +author: John Snow Labs +name: t5_grammar_model18000 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_grammar_model18000` is a English model originally trained by kalobiralo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_grammar_model18000_en_5.4.2_3.0_1723940331428.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_grammar_model18000_en_5.4.2_3.0_1723940331428.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_grammar_model18000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_grammar_model18000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_grammar_model18000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/kalobiralo/t5-grammar-model18000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_pipeline_en.md new file mode 100644 index 00000000000000..1427cebb291c29 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_grammar_model18000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_grammar_model18000_pipeline pipeline T5Transformer from kalobiralo +author: John Snow Labs +name: t5_grammar_model18000_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_grammar_model18000_pipeline` is a English model originally trained by kalobiralo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_grammar_model18000_pipeline_en_5.4.2_3.0_1723940380665.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_grammar_model18000_pipeline_en_5.4.2_3.0_1723940380665.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_grammar_model18000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_grammar_model18000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_grammar_model18000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/kalobiralo/t5-grammar-model18000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_ja.md new file mode 100644 index 00000000000000..ec010f0279fbae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_ja.md @@ -0,0 +1,93 @@ +--- +layout: model +title: Japanese T5ForConditionalGeneration Cased model (from astremo) +author: John Snow Labs +name: t5_jainu +date: 2024-08-18 +tags: [ja, open_source, t5, onnx] +task: Text Generation +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `JAINU` is a Japanese model originally trained by `astremo`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jainu_ja_5.4.2_3.0_1724010889975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jainu_ja_5.4.2_3.0_1724010889975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_jainu","ja") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_jainu","ja") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jainu| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|1.0 GB| + +## References + +References + +- https://huggingface.co/astremo/JAINU +- http://creativecommons.org/licenses/by/4.0/ +- http://creativecommons.org/licenses/by/4.0/ +- http://creativecommons.org/licenses/by/4.0/ \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_pipeline_ja.md b/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_pipeline_ja.md new file mode 100644 index 00000000000000..664450e2aa1edd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_jainu_pipeline_ja.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Japanese t5_jainu_pipeline pipeline T5Transformer from astremo +author: John Snow Labs +name: t5_jainu_pipeline +date: 2024-08-18 +tags: [ja, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_jainu_pipeline` is a Japanese model originally trained by astremo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jainu_pipeline_ja_5.4.2_3.0_1724010937725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jainu_pipeline_ja_5.4.2_3.0_1724010937725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_jainu_pipeline", lang = "ja") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_jainu_pipeline", lang = "ja") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jainu_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ja| +|Size:|1.0 GB| + +## References + +https://huggingface.co/astremo/JAINU + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_en.md new file mode 100644 index 00000000000000..e38b10405a0225 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_jump_commands_tonga_tonga_islands_actions_5 T5Transformer from LadyShizu +author: John Snow Labs +name: t5_jump_commands_tonga_tonga_islands_actions_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_jump_commands_tonga_tonga_islands_actions_5` is a English model originally trained by LadyShizu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jump_commands_tonga_tonga_islands_actions_5_en_5.4.2_3.0_1723982382752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jump_commands_tonga_tonga_islands_actions_5_en_5.4.2_3.0_1723982382752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_jump_commands_tonga_tonga_islands_actions_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_jump_commands_tonga_tonga_islands_actions_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jump_commands_tonga_tonga_islands_actions_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LadyShizu/T5_jump-commands_to_actions_5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_pipeline_en.md new file mode 100644 index 00000000000000..76639eff1273bc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_jump_commands_tonga_tonga_islands_actions_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_jump_commands_tonga_tonga_islands_actions_5_pipeline pipeline T5Transformer from LadyShizu +author: John Snow Labs +name: t5_jump_commands_tonga_tonga_islands_actions_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_jump_commands_tonga_tonga_islands_actions_5_pipeline` is a English model originally trained by LadyShizu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_jump_commands_tonga_tonga_islands_actions_5_pipeline_en_5.4.2_3.0_1723982436678.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_jump_commands_tonga_tonga_islands_actions_5_pipeline_en_5.4.2_3.0_1723982436678.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_jump_commands_tonga_tonga_islands_actions_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_jump_commands_tonga_tonga_islands_actions_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_jump_commands_tonga_tonga_islands_actions_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LadyShizu/T5_jump-commands_to_actions_5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_en.md new file mode 100644 index 00000000000000..b5da775128d4d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_en.md @@ -0,0 +1,90 @@ +--- +layout: model +title: English T5ForConditionalGeneration Base Cased model (from swcrazyfan) +author: John Snow Labs +name: t5_kingjamesify_base +date: 2024-08-18 +tags: [en, open_source, t5, onnx] +task: Text Generation +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `KingJamesify-T5-Base` is a English model originally trained by `swcrazyfan`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_kingjamesify_base_en_5.4.2_3.0_1724011414121.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_kingjamesify_base_en_5.4.2_3.0_1724011414121.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_kingjamesify_base","en") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_kingjamesify_base","en") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_kingjamesify_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +References + +- https://huggingface.co/swcrazyfan/KingJamesify-T5-Base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_pipeline_en.md new file mode 100644 index 00000000000000..b0414e50d9b58d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_kingjamesify_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_kingjamesify_base_pipeline pipeline T5Transformer from swcrazyfan +author: John Snow Labs +name: t5_kingjamesify_base_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_kingjamesify_base_pipeline` is a English model originally trained by swcrazyfan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_kingjamesify_base_pipeline_en_5.4.2_3.0_1724011462927.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_kingjamesify_base_pipeline_en_5.4.2_3.0_1724011462927.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_kingjamesify_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_kingjamesify_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_kingjamesify_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/swcrazyfan/KingJamesify-T5-Base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_en.md new file mode 100644 index 00000000000000..5dc1ede28f03d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_coedit T5Transformer from iliazlobin +author: John Snow Labs +name: t5_large_coedit +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_coedit` is a English model originally trained by iliazlobin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_coedit_en_5.4.2_3.0_1723946976145.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_coedit_en_5.4.2_3.0_1723946976145.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_coedit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_coedit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_coedit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/iliazlobin/t5-large-coedit \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_pipeline_en.md new file mode 100644 index 00000000000000..6c1b294bbd1e2f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_coedit_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_coedit_pipeline pipeline T5Transformer from iliazlobin +author: John Snow Labs +name: t5_large_coedit_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_coedit_pipeline` is a English model originally trained by iliazlobin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_coedit_pipeline_en_5.4.2_3.0_1723947126315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_coedit_pipeline_en_5.4.2_3.0_1723947126315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_coedit_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_coedit_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_coedit_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/iliazlobin/t5-large-coedit + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_fine_tune_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_fine_tune_en.md new file mode 100644 index 00000000000000..f67719a4e3f74c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_fine_tune_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_fine_tune T5Transformer from baesad +author: John Snow Labs +name: t5_large_fine_tune +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_fine_tune` is a English model originally trained by baesad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_fine_tune_en_5.4.2_3.0_1723942658284.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_fine_tune_en_5.4.2_3.0_1723942658284.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_fine_tune","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_fine_tune", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_fine_tune| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/baesad/t5-large-fine-tune \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_finetuned_summarize_scientific_articles_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_finetuned_summarize_scientific_articles_en.md new file mode 100644 index 00000000000000..ee2e54c47faf97 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_finetuned_summarize_scientific_articles_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_finetuned_summarize_scientific_articles T5Transformer from sana-ngu +author: John Snow Labs +name: t5_large_finetuned_summarize_scientific_articles +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_finetuned_summarize_scientific_articles` is a English model originally trained by sana-ngu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_summarize_scientific_articles_en_5.4.2_3.0_1723957175907.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_summarize_scientific_articles_en_5.4.2_3.0_1723957175907.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_finetuned_summarize_scientific_articles","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_finetuned_summarize_scientific_articles", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_finetuned_summarize_scientific_articles| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/sana-ngu/t5-large-finetuned-summarize-scientific-articles \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_en.md new file mode 100644 index 00000000000000..b1329f1fa05c6d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_headline_generator_testing_1 T5Transformer from abdulmatinomotoso +author: John Snow Labs +name: t5_large_headline_generator_testing_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_headline_generator_testing_1` is a English model originally trained by abdulmatinomotoso. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_headline_generator_testing_1_en_5.4.2_3.0_1723986507609.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_headline_generator_testing_1_en_5.4.2_3.0_1723986507609.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_headline_generator_testing_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_headline_generator_testing_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_headline_generator_testing_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/abdulmatinomotoso/t5_large_headline_generator_testing_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_pipeline_en.md new file mode 100644 index 00000000000000..7dda3ccd18af5b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_headline_generator_testing_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_headline_generator_testing_1_pipeline pipeline T5Transformer from abdulmatinomotoso +author: John Snow Labs +name: t5_large_headline_generator_testing_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_headline_generator_testing_1_pipeline` is a English model originally trained by abdulmatinomotoso. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_headline_generator_testing_1_pipeline_en_5.4.2_3.0_1723986639088.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_headline_generator_testing_1_pipeline_en_5.4.2_3.0_1723986639088.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_headline_generator_testing_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_headline_generator_testing_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_headline_generator_testing_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/abdulmatinomotoso/t5_large_headline_generator_testing_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_en.md new file mode 100644 index 00000000000000..3cde7b813910f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_radiology_ai_imagingcancer_0_9 T5Transformer from hisaoka +author: John Snow Labs +name: t5_large_radiology_ai_imagingcancer_0_9 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_radiology_ai_imagingcancer_0_9` is a English model originally trained by hisaoka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_9_en_5.4.2_3.0_1724006158806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_9_en_5.4.2_3.0_1724006158806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_radiology_ai_imagingcancer_0_9","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_radiology_ai_imagingcancer_0_9", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_radiology_ai_imagingcancer_0_9| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/hisaoka/t5-large_radiology-ai-imagingcancer-0.9 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_pipeline_en.md new file mode 100644 index 00000000000000..b3e5b8408143ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_radiology_ai_imagingcancer_0_9_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_radiology_ai_imagingcancer_0_9_pipeline pipeline T5Transformer from hisaoka +author: John Snow Labs +name: t5_large_radiology_ai_imagingcancer_0_9_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_radiology_ai_imagingcancer_0_9_pipeline` is a English model originally trained by hisaoka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_9_pipeline_en_5.4.2_3.0_1724006308823.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_9_pipeline_en_5.4.2_3.0_1724006308823.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_radiology_ai_imagingcancer_0_9_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_radiology_ai_imagingcancer_0_9_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_radiology_ai_imagingcancer_0_9_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/hisaoka/t5-large_radiology-ai-imagingcancer-0.9 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_squad_ae_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_squad_ae_pipeline_en.md new file mode 100644 index 00000000000000..3b626a33315d34 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_squad_ae_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_squad_ae_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: t5_large_squad_ae_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_squad_ae_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_squad_ae_pipeline_en_5.4.2_3.0_1724017220758.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_squad_ae_pipeline_en_5.4.2_3.0_1724017220758.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_squad_ae_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_squad_ae_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_squad_ae_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/lmqg/t5-large-squad-ae + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_en.md new file mode 100644 index 00000000000000..c82ae0c5209553 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_wikisql T5Transformer from nannullna +author: John Snow Labs +name: t5_large_wikisql +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_wikisql` is a English model originally trained by nannullna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_wikisql_en_5.4.2_3.0_1724003097228.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_wikisql_en_5.4.2_3.0_1724003097228.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_wikisql","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_wikisql", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_wikisql| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/nannullna/t5-large-wikisql \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_pipeline_en.md new file mode 100644 index 00000000000000..ffe869eeeb1833 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_large_wikisql_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_wikisql_pipeline pipeline T5Transformer from nannullna +author: John Snow Labs +name: t5_large_wikisql_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_wikisql_pipeline` is a English model originally trained by nannullna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_wikisql_pipeline_en_5.4.2_3.0_1724003231688.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_wikisql_pipeline_en_5.4.2_3.0_1724003231688.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_wikisql_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_wikisql_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_wikisql_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/nannullna/t5-large-wikisql + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_en.md new file mode 100644 index 00000000000000..a13e769b70a4f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_med_ocr_aug_small T5Transformer from lowem1 +author: John Snow Labs +name: t5_med_ocr_aug_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_med_ocr_aug_small` is a English model originally trained by lowem1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_med_ocr_aug_small_en_5.4.2_3.0_1724000649215.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_med_ocr_aug_small_en_5.4.2_3.0_1724000649215.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_med_ocr_aug_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_med_ocr_aug_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_med_ocr_aug_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|314.8 MB| + +## References + +https://huggingface.co/lowem1/t5_med_ocr_aug-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_pipeline_en.md new file mode 100644 index 00000000000000..0b46614e8873c3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_med_ocr_aug_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_med_ocr_aug_small_pipeline pipeline T5Transformer from lowem1 +author: John Snow Labs +name: t5_med_ocr_aug_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_med_ocr_aug_small_pipeline` is a English model originally trained by lowem1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_med_ocr_aug_small_pipeline_en_5.4.2_3.0_1724000673640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_med_ocr_aug_small_pipeline_en_5.4.2_3.0_1724000673640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_med_ocr_aug_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_med_ocr_aug_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_med_ocr_aug_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|314.8 MB| + +## References + +https://huggingface.co/lowem1/t5_med_ocr_aug-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_en.md new file mode 100644 index 00000000000000..acac597352ae76 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_model_1_d_4 T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_d_4 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_d_4` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_d_4_en_5.4.2_3.0_1723992839956.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_d_4_en_5.4.2_3.0_1723992839956.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_model_1_d_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_model_1_d_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_d_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|963.2 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-d-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_pipeline_en.md new file mode 100644 index 00000000000000..c1dc973a663acb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_d_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_model_1_d_4_pipeline pipeline T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_d_4_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_d_4_pipeline` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_d_4_pipeline_en_5.4.2_3.0_1723992888175.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_d_4_pipeline_en_5.4.2_3.0_1723992888175.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_model_1_d_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_model_1_d_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_d_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|963.2 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-d-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_en.md new file mode 100644 index 00000000000000..ab565822451b85 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_model_1_feedback_e1 T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_feedback_e1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_feedback_e1` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_e1_en_5.4.2_3.0_1723950415001.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_e1_en_5.4.2_3.0_1723950415001.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_model_1_feedback_e1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_model_1_feedback_e1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_feedback_e1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|965.1 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-feedback-e1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_pipeline_en.md new file mode 100644 index 00000000000000..81223a51f32a7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_model_1_feedback_e1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_model_1_feedback_e1_pipeline pipeline T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_feedback_e1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_feedback_e1_pipeline` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_e1_pipeline_en_5.4.2_3.0_1723950466076.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_e1_pipeline_en_5.4.2_3.0_1723950466076.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_model_1_feedback_e1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_model_1_feedback_e1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_feedback_e1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|965.1 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-feedback-e1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_de.md b/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_de.md new file mode 100644 index 00000000000000..3bdf28de151f6a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_de.md @@ -0,0 +1,91 @@ +--- +layout: model +title: German T5ForConditionalGeneration Small Cased model (from ml6team) +author: John Snow Labs +name: t5_mt5_small_german_finetune_mlsum +date: 2024-08-18 +tags: [de, open_source, t5, onnx] +task: Text Generation +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `mt5-small-german-finetune-mlsum` is a German model originally trained by `ml6team`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_mt5_small_german_finetune_mlsum_de_5.4.2_3.0_1724010367051.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_mt5_small_german_finetune_mlsum_de_5.4.2_3.0_1724010367051.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_mt5_small_german_finetune_mlsum","de") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_mt5_small_german_finetune_mlsum","de") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_mt5_small_german_finetune_mlsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|de| +|Size:|1.3 GB| + +## References + +References + +- https://huggingface.co/ml6team/mt5-small-german-finetune-mlsum +- https://github.com/pltrdy/rouge \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_pipeline_de.md b/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_pipeline_de.md new file mode 100644 index 00000000000000..1d1a1824f15179 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_mt5_small_german_finetune_mlsum_pipeline_de.md @@ -0,0 +1,69 @@ +--- +layout: model +title: German t5_mt5_small_german_finetune_mlsum_pipeline pipeline T5Transformer from ml6team +author: John Snow Labs +name: t5_mt5_small_german_finetune_mlsum_pipeline +date: 2024-08-18 +tags: [de, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: de +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_mt5_small_german_finetune_mlsum_pipeline` is a German model originally trained by ml6team. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_mt5_small_german_finetune_mlsum_pipeline_de_5.4.2_3.0_1724010515641.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_mt5_small_german_finetune_mlsum_pipeline_de_5.4.2_3.0_1724010515641.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_mt5_small_german_finetune_mlsum_pipeline", lang = "de") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_mt5_small_german_finetune_mlsum_pipeline", lang = "de") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_mt5_small_german_finetune_mlsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|de| +|Size:|1.3 GB| + +## References + +https://huggingface.co/ml6team/mt5-small-german-finetune-mlsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_protect_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_protect_en.md new file mode 100644 index 00000000000000..ab4f0daff9e387 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_protect_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_protect T5Transformer from AnonymousSubmissionOnly +author: John Snow Labs +name: t5_protect +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_protect` is a English model originally trained by AnonymousSubmissionOnly. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_protect_en_5.4.2_3.0_1723950081866.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_protect_en_5.4.2_3.0_1723950081866.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_protect","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_protect", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_protect| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/AnonymousSubmissionOnly/t5-protect \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_protect_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_protect_pipeline_en.md new file mode 100644 index 00000000000000..6efa74c5ba6d1e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_protect_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_protect_pipeline pipeline T5Transformer from AnonymousSubmissionOnly +author: John Snow Labs +name: t5_protect_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_protect_pipeline` is a English model originally trained by AnonymousSubmissionOnly. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_protect_pipeline_en_5.4.2_3.0_1723950137581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_protect_pipeline_en_5.4.2_3.0_1723950137581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_protect_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_protect_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_protect_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/AnonymousSubmissionOnly/t5-protect + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_en.md new file mode 100644 index 00000000000000..99a836c3a2d61c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_recommendation_sports_equipment_english_diptipawar T5Transformer from DiptiPawar +author: John Snow Labs +name: t5_recommendation_sports_equipment_english_diptipawar +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_recommendation_sports_equipment_english_diptipawar` is a English model originally trained by DiptiPawar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_diptipawar_en_5.4.2_3.0_1723949449695.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_diptipawar_en_5.4.2_3.0_1723949449695.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_recommendation_sports_equipment_english_diptipawar","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_recommendation_sports_equipment_english_diptipawar", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_recommendation_sports_equipment_english_diptipawar| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.7 GB| + +## References + +https://huggingface.co/DiptiPawar/t5_recommendation_sports_equipment_english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_pipeline_en.md new file mode 100644 index 00000000000000..a483863690230a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_recommendation_sports_equipment_english_diptipawar_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_recommendation_sports_equipment_english_diptipawar_pipeline pipeline T5Transformer from DiptiPawar +author: John Snow Labs +name: t5_recommendation_sports_equipment_english_diptipawar_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_recommendation_sports_equipment_english_diptipawar_pipeline` is a English model originally trained by DiptiPawar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_diptipawar_pipeline_en_5.4.2_3.0_1723949675748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_diptipawar_pipeline_en_5.4.2_3.0_1723949675748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_recommendation_sports_equipment_english_diptipawar_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_recommendation_sports_equipment_english_diptipawar_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_recommendation_sports_equipment_english_diptipawar_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.7 GB| + +## References + +https://huggingface.co/DiptiPawar/t5_recommendation_sports_equipment_english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_en.md new file mode 100644 index 00000000000000..59fea298329c2b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_regex_summarization T5Transformer from rbesaleli +author: John Snow Labs +name: t5_regex_summarization +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_regex_summarization` is a English model originally trained by rbesaleli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_regex_summarization_en_5.4.2_3.0_1724024113035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_regex_summarization_en_5.4.2_3.0_1724024113035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_regex_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_regex_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_regex_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/rbesaleli/t5-regex-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_pipeline_en.md new file mode 100644 index 00000000000000..5dc17c186e8e12 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_regex_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_regex_summarization_pipeline pipeline T5Transformer from rbesaleli +author: John Snow Labs +name: t5_regex_summarization_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_regex_summarization_pipeline` is a English model originally trained by rbesaleli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_regex_summarization_pipeline_en_5.4.2_3.0_1724024167964.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_regex_summarization_pipeline_en_5.4.2_3.0_1724024167964.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_regex_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_regex_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_regex_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/rbesaleli/t5-regex-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_en.md new file mode 100644 index 00000000000000..d7e0b0b727f012 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_samsum_jvelja T5Transformer from jvelja +author: John Snow Labs +name: t5_samsum_jvelja +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_samsum_jvelja` is a English model originally trained by jvelja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_samsum_jvelja_en_5.4.2_3.0_1724025194505.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_samsum_jvelja_en_5.4.2_3.0_1724025194505.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_samsum_jvelja","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_samsum_jvelja", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_samsum_jvelja| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jvelja/t5-samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_pipeline_en.md new file mode 100644 index 00000000000000..8f6d7c438633b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_samsum_jvelja_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_samsum_jvelja_pipeline pipeline T5Transformer from jvelja +author: John Snow Labs +name: t5_samsum_jvelja_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_samsum_jvelja_pipeline` is a English model originally trained by jvelja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_samsum_jvelja_pipeline_en_5.4.2_3.0_1724025327774.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_samsum_jvelja_pipeline_en_5.4.2_3.0_1724025327774.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_samsum_jvelja_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_samsum_jvelja_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_samsum_jvelja_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jvelja/t5-samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_en.md new file mode 100644 index 00000000000000..6819b1328c138c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_sentence_paraphraser T5Transformer from ramsrigouthamg +author: John Snow Labs +name: t5_sentence_paraphraser +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_sentence_paraphraser` is a English model originally trained by ramsrigouthamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_sentence_paraphraser_en_5.4.2_3.0_1724012926672.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_sentence_paraphraser_en_5.4.2_3.0_1724012926672.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_sentence_paraphraser","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_sentence_paraphraser", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_sentence_paraphraser| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ramsrigouthamg/t5_sentence_paraphraser \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_pipeline_en.md new file mode 100644 index 00000000000000..82dbe2c6e44179 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_sentence_paraphraser_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_sentence_paraphraser_pipeline pipeline T5Transformer from ramsrigouthamg +author: John Snow Labs +name: t5_sentence_paraphraser_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_sentence_paraphraser_pipeline` is a English model originally trained by ramsrigouthamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_sentence_paraphraser_pipeline_en_5.4.2_3.0_1724012973995.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_sentence_paraphraser_pipeline_en_5.4.2_3.0_1724012973995.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_sentence_paraphraser_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_sentence_paraphraser_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_sentence_paraphraser_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ramsrigouthamg/t5_sentence_paraphraser + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_en.md new file mode 100644 index 00000000000000..3ed21849f30b07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_act2pas_pushparaj20 T5Transformer from Pushparaj20 +author: John Snow Labs +name: t5_small_act2pas_pushparaj20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_act2pas_pushparaj20` is a English model originally trained by Pushparaj20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_act2pas_pushparaj20_en_5.4.2_3.0_1723987872053.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_act2pas_pushparaj20_en_5.4.2_3.0_1723987872053.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_act2pas_pushparaj20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_act2pas_pushparaj20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_act2pas_pushparaj20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.8 MB| + +## References + +https://huggingface.co/Pushparaj20/t5-small-act2pas \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_pipeline_en.md new file mode 100644 index 00000000000000..315457a5837909 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_act2pas_pushparaj20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_act2pas_pushparaj20_pipeline pipeline T5Transformer from Pushparaj20 +author: John Snow Labs +name: t5_small_act2pas_pushparaj20_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_act2pas_pushparaj20_pipeline` is a English model originally trained by Pushparaj20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_act2pas_pushparaj20_pipeline_en_5.4.2_3.0_1723987889165.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_act2pas_pushparaj20_pipeline_en_5.4.2_3.0_1723987889165.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_act2pas_pushparaj20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_act2pas_pushparaj20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_act2pas_pushparaj20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.8 MB| + +## References + +https://huggingface.co/Pushparaj20/t5-small-act2pas + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_demosthene_or_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_demosthene_or_pipeline_en.md new file mode 100644 index 00000000000000..342d8b02f584cd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_demosthene_or_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_demosthene_or_pipeline pipeline T5Transformer from Demosthene-OR +author: John Snow Labs +name: t5_small_demosthene_or_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_demosthene_or_pipeline` is a English model originally trained by Demosthene-OR. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_demosthene_or_pipeline_en_5.4.2_3.0_1723955708234.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_demosthene_or_pipeline_en_5.4.2_3.0_1723955708234.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_demosthene_or_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_demosthene_or_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_demosthene_or_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|953.9 MB| + +## References + +https://huggingface.co/Demosthene-OR/t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_en.md new file mode 100644 index 00000000000000..ecb44713a97d62 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_few_shot_k_64_finetuned_squad_seed_2 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_64_finetuned_squad_seed_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_64_finetuned_squad_seed_2` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_64_finetuned_squad_seed_2_en_5.4.2_3.0_1723953988955.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_64_finetuned_squad_seed_2_en_5.4.2_3.0_1723953988955.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_few_shot_k_64_finetuned_squad_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_few_shot_k_64_finetuned_squad_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_64_finetuned_squad_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|297.5 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-64-finetuned-squad-seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..5e8f0d7bc1401c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline_en_5.4.2_3.0_1723954019954.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline_en_5.4.2_3.0_1723954019954.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_64_finetuned_squad_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|297.5 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-64-finetuned-squad-seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_en.md new file mode 100644 index 00000000000000..dc539f4e7fc269 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetune_ag_news_scratch T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_small_finetune_ag_news_scratch +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetune_ag_news_scratch` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetune_ag_news_scratch_en_5.4.2_3.0_1723990729744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetune_ag_news_scratch_en_5.4.2_3.0_1723990729744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetune_ag_news_scratch","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetune_ag_news_scratch", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetune_ag_news_scratch| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_small_finetune_ag_news_scratch \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_pipeline_en.md new file mode 100644 index 00000000000000..72865a1257abc5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetune_ag_news_scratch_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetune_ag_news_scratch_pipeline pipeline T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_small_finetune_ag_news_scratch_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetune_ag_news_scratch_pipeline` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetune_ag_news_scratch_pipeline_en_5.4.2_3.0_1723990748599.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetune_ag_news_scratch_pipeline_en_5.4.2_3.0_1723990748599.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetune_ag_news_scratch_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetune_ag_news_scratch_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetune_ag_news_scratch_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_small_finetune_ag_news_scratch + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_en.md new file mode 100644 index 00000000000000..0b58839f00311b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_car_dataset_shangari T5Transformer from shangari +author: John Snow Labs +name: t5_small_finetuned_car_dataset_shangari +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_car_dataset_shangari` is a English model originally trained by shangari. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_car_dataset_shangari_en_5.4.2_3.0_1723995080321.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_car_dataset_shangari_en_5.4.2_3.0_1723995080321.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_car_dataset_shangari","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_car_dataset_shangari", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_car_dataset_shangari| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|319.8 MB| + +## References + +https://huggingface.co/shangari/t5-small-finetuned-car_dataset \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_pipeline_en.md new file mode 100644 index 00000000000000..21f5ca5c044c42 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_car_dataset_shangari_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_car_dataset_shangari_pipeline pipeline T5Transformer from shangari +author: John Snow Labs +name: t5_small_finetuned_car_dataset_shangari_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_car_dataset_shangari_pipeline` is a English model originally trained by shangari. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_car_dataset_shangari_pipeline_en_5.4.2_3.0_1723995098243.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_car_dataset_shangari_pipeline_en_5.4.2_3.0_1723995098243.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_car_dataset_shangari_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_car_dataset_shangari_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_car_dataset_shangari_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|319.8 MB| + +## References + +https://huggingface.co/shangari/t5-small-finetuned-car_dataset + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_en.md new file mode 100644 index 00000000000000..4fc1b59b246ad0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99 T5Transformer from mohit-99 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99` is a English model originally trained by mohit-99. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_en_5.4.2_3.0_1723989241461.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_en_5.4.2_3.0_1723989241461.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.7 MB| + +## References + +https://huggingface.co/mohit-99/t5-small-finetuned-en-to-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline_en.md new file mode 100644 index 00000000000000..752dd824c61c40 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline pipeline T5Transformer from mohit-99 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline` is a English model originally trained by mohit-99. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline_en_5.4.2_3.0_1723989260524.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline_en_5.4.2_3.0_1723989260524.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_german_mohit_99_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.7 MB| + +## References + +https://huggingface.co/mohit-99/t5-small-finetuned-en-to-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_en.md new file mode 100644 index 00000000000000..f68ad6803f8eb0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_italian_hrs T5Transformer from din0s +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_italian_hrs +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_italian_hrs` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_en_5.4.2_3.0_1723994063856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_en_5.4.2_3.0_1723994063856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_italian_hrs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_italian_hrs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_italian_hrs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/din0s/t5-small-finetuned-en-to-it-hrs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline_en.md new file mode 100644 index 00000000000000..6f9dec8a48454b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline pipeline T5Transformer from din0s +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline_en_5.4.2_3.0_1723994079707.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline_en_5.4.2_3.0_1723994079707.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_italian_hrs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/din0s/t5-small-finetuned-en-to-it-hrs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_en.md new file mode 100644 index 00000000000000..671b53734385f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_german_tonga_tonga_islands_english_swd T5Transformer from marciovbarbosa +author: John Snow Labs +name: t5_small_finetuned_german_tonga_tonga_islands_english_swd +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_tonga_tonga_islands_english_swd` is a English model originally trained by marciovbarbosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_swd_en_5.4.2_3.0_1723987016897.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_swd_en_5.4.2_3.0_1723987016897.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_german_tonga_tonga_islands_english_swd","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_german_tonga_tonga_islands_english_swd", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_tonga_tonga_islands_english_swd| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.8 MB| + +## References + +https://huggingface.co/marciovbarbosa/t5-small-finetuned-de-to-en-swd \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline_en.md new file mode 100644 index 00000000000000..d9e63016959a92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline pipeline T5Transformer from marciovbarbosa +author: John Snow Labs +name: t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline` is a English model originally trained by marciovbarbosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline_en_5.4.2_3.0_1723987039901.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline_en_5.4.2_3.0_1723987039901.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_tonga_tonga_islands_english_swd_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.8 MB| + +## References + +https://huggingface.co/marciovbarbosa/t5-small-finetuned-de-to-en-swd + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_en.md new file mode 100644 index 00000000000000..eddb030c65f5b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_orca T5Transformer from Megareyka +author: John Snow Labs +name: t5_small_finetuned_orca +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_orca` is a English model originally trained by Megareyka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_orca_en_5.4.2_3.0_1724000760247.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_orca_en_5.4.2_3.0_1724000760247.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_orca","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_orca", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_orca| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.1 MB| + +## References + +https://huggingface.co/Megareyka/t5-small-finetuned-Orca \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_pipeline_en.md new file mode 100644 index 00000000000000..56a33f194dde5a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_orca_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_orca_pipeline pipeline T5Transformer from Megareyka +author: John Snow Labs +name: t5_small_finetuned_orca_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_orca_pipeline` is a English model originally trained by Megareyka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_orca_pipeline_en_5.4.2_3.0_1724000778003.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_orca_pipeline_en_5.4.2_3.0_1724000778003.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_orca_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_orca_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_orca_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.1 MB| + +## References + +https://huggingface.co/Megareyka/t5-small-finetuned-Orca + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_en.md new file mode 100644 index 00000000000000..21d247394ae5d3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_pushparaj20 T5Transformer from Pushparaj20 +author: John Snow Labs +name: t5_small_finetuned_pushparaj20 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_pushparaj20` is a English model originally trained by Pushparaj20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_pushparaj20_en_5.4.2_3.0_1723999568536.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_pushparaj20_en_5.4.2_3.0_1723999568536.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_pushparaj20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_pushparaj20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_pushparaj20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.7 MB| + +## References + +https://huggingface.co/Pushparaj20/t5-small-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_pipeline_en.md new file mode 100644 index 00000000000000..02198ffe727cbb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_pushparaj20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_pushparaj20_pipeline pipeline T5Transformer from Pushparaj20 +author: John Snow Labs +name: t5_small_finetuned_pushparaj20_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_pushparaj20_pipeline` is a English model originally trained by Pushparaj20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_pushparaj20_pipeline_en_5.4.2_3.0_1723999586352.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_pushparaj20_pipeline_en_5.4.2_3.0_1723999586352.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_pushparaj20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_pushparaj20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_pushparaj20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.7 MB| + +## References + +https://huggingface.co/Pushparaj20/t5-small-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_en.md new file mode 100644 index 00000000000000..05ef7e53a3e0fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_sci_tldr T5Transformer from SunshineYellow +author: John Snow Labs +name: t5_small_finetuned_sci_tldr +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_sci_tldr` is a English model originally trained by SunshineYellow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_sci_tldr_en_5.4.2_3.0_1723972768643.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_sci_tldr_en_5.4.2_3.0_1723972768643.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_sci_tldr","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_sci_tldr", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_sci_tldr| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.9 MB| + +## References + +https://huggingface.co/SunshineYellow/t5-small-finetuned-sci_tldr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_pipeline_en.md new file mode 100644 index 00000000000000..aa62309737a5c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_sci_tldr_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_sci_tldr_pipeline pipeline T5Transformer from SunshineYellow +author: John Snow Labs +name: t5_small_finetuned_sci_tldr_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_sci_tldr_pipeline` is a English model originally trained by SunshineYellow. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_sci_tldr_pipeline_en_5.4.2_3.0_1723972785566.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_sci_tldr_pipeline_en_5.4.2_3.0_1723972785566.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_sci_tldr_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_sci_tldr_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_sci_tldr_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.9 MB| + +## References + +https://huggingface.co/SunshineYellow/t5-small-finetuned-sci_tldr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_en.md new file mode 100644 index 00000000000000..81cf2ee646ea5c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_squad_infilling_lr_1e_4 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_finetuned_squad_infilling_lr_1e_4 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_squad_infilling_lr_1e_4` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_squad_infilling_lr_1e_4_en_5.4.2_3.0_1723989323692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_squad_infilling_lr_1e_4_en_5.4.2_3.0_1723989323692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_squad_infilling_lr_1e_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_squad_infilling_lr_1e_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_squad_infilling_lr_1e_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|337.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-finetuned-squad-infilling-lr-1e-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_pipeline_en.md new file mode 100644 index 00000000000000..15c94fe1644bf5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_squad_infilling_lr_1e_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_squad_infilling_lr_1e_4_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_finetuned_squad_infilling_lr_1e_4_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_squad_infilling_lr_1e_4_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_squad_infilling_lr_1e_4_pipeline_en_5.4.2_3.0_1723989342309.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_squad_infilling_lr_1e_4_pipeline_en_5.4.2_3.0_1723989342309.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_squad_infilling_lr_1e_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_squad_infilling_lr_1e_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_squad_infilling_lr_1e_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|337.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-finetuned-squad-infilling-lr-1e-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_en.md new file mode 100644 index 00000000000000..a5df172942bf0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol T5Transformer from anki08 +author: John Snow Labs +name: t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol` is a English model originally trained by anki08. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_en_5.4.2_3.0_1723955722744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_en_5.4.2_3.0_1723955722744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|347.7 MB| + +## References + +https://huggingface.co/anki08/t5-small-finetuned-text2log-finetuned-nl-to-fol \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline_en.md new file mode 100644 index 00000000000000..eceaff303e204b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline pipeline T5Transformer from anki08 +author: John Snow Labs +name: t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline` is a English model originally trained by anki08. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline_en_5.4.2_3.0_1723955739819.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline_en_5.4.2_3.0_1723955739819.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_text2log_finetuned_dutch_tonga_tonga_islands_fol_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|347.7 MB| + +## References + +https://huggingface.co/anki08/t5-small-finetuned-text2log-finetuned-nl-to-fol + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_en.md new file mode 100644 index 00000000000000..aeffa004e34be3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_chaoyivision T5Transformer from chaoyivision +author: John Snow Labs +name: t5_small_finetuned_xsum_chaoyivision +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_chaoyivision` is a English model originally trained by chaoyivision. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_chaoyivision_en_5.4.2_3.0_1723965676202.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_chaoyivision_en_5.4.2_3.0_1723965676202.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_chaoyivision","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_chaoyivision", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_chaoyivision| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|838.5 KB| + +## References + +https://huggingface.co/chaoyivision/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_pipeline_en.md new file mode 100644 index 00000000000000..28aab2e551bf3c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_chaoyivision_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_chaoyivision_pipeline pipeline T5Transformer from chaoyivision +author: John Snow Labs +name: t5_small_finetuned_xsum_chaoyivision_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_chaoyivision_pipeline` is a English model originally trained by chaoyivision. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_chaoyivision_pipeline_en_5.4.2_3.0_1723965677610.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_chaoyivision_pipeline_en_5.4.2_3.0_1723965677610.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_chaoyivision_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_chaoyivision_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_chaoyivision_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|841.7 KB| + +## References + +https://huggingface.co/chaoyivision/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_en.md new file mode 100644 index 00000000000000..675b06acbba702 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_datasteves T5Transformer from DataSteves +author: John Snow Labs +name: t5_small_finetuned_xsum_datasteves +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_datasteves` is a English model originally trained by DataSteves. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_datasteves_en_5.4.2_3.0_1723949212638.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_datasteves_en_5.4.2_3.0_1723949212638.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_datasteves","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_datasteves", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_datasteves| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.5 MB| + +## References + +https://huggingface.co/DataSteves/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_pipeline_en.md new file mode 100644 index 00000000000000..9aab6323031737 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_datasteves_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_datasteves_pipeline pipeline T5Transformer from DataSteves +author: John Snow Labs +name: t5_small_finetuned_xsum_datasteves_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_datasteves_pipeline` is a English model originally trained by DataSteves. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_datasteves_pipeline_en_5.4.2_3.0_1723949230227.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_datasteves_pipeline_en_5.4.2_3.0_1723949230227.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_datasteves_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_datasteves_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_datasteves_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.5 MB| + +## References + +https://huggingface.co/DataSteves/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_en.md new file mode 100644 index 00000000000000..0ac6f40c4cf703 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_jpodivin T5Transformer from jpodivin +author: John Snow Labs +name: t5_small_finetuned_xsum_jpodivin +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_jpodivin` is a English model originally trained by jpodivin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_jpodivin_en_5.4.2_3.0_1723953143991.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_jpodivin_en_5.4.2_3.0_1723953143991.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_jpodivin","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_jpodivin", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_jpodivin| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|321.0 MB| + +## References + +https://huggingface.co/jpodivin/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_pipeline_en.md new file mode 100644 index 00000000000000..0be9d5135a6f41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_jpodivin_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_jpodivin_pipeline pipeline T5Transformer from jpodivin +author: John Snow Labs +name: t5_small_finetuned_xsum_jpodivin_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_jpodivin_pipeline` is a English model originally trained by jpodivin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_jpodivin_pipeline_en_5.4.2_3.0_1723953165683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_jpodivin_pipeline_en_5.4.2_3.0_1723953165683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_jpodivin_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_jpodivin_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_jpodivin_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|321.0 MB| + +## References + +https://huggingface.co/jpodivin/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_en.md new file mode 100644 index 00000000000000..e785bf9193f68c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_purpose_system T5Transformer from fanzru +author: John Snow Labs +name: t5_small_finetuned_xsum_purpose_system +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_purpose_system` is a English model originally trained by fanzru. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_purpose_system_en_5.4.2_3.0_1723969569472.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_purpose_system_en_5.4.2_3.0_1723969569472.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_purpose_system","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_purpose_system", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_purpose_system| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.6 MB| + +## References + +https://huggingface.co/fanzru/t5-small-finetuned-xsum-purpose-system \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_pipeline_en.md new file mode 100644 index 00000000000000..7e0d0ec98c54a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_finetuned_xsum_purpose_system_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_purpose_system_pipeline pipeline T5Transformer from fanzru +author: John Snow Labs +name: t5_small_finetuned_xsum_purpose_system_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_purpose_system_pipeline` is a English model originally trained by fanzru. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_purpose_system_pipeline_en_5.4.2_3.0_1723969586858.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_purpose_system_pipeline_en_5.4.2_3.0_1723969586858.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_purpose_system_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_purpose_system_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_purpose_system_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.6 MB| + +## References + +https://huggingface.co/fanzru/t5-small-finetuned-xsum-purpose-system + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_en.md new file mode 100644 index 00000000000000..ac13532b001f52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_french_title_generation T5Transformer from edgolyakova +author: John Snow Labs +name: t5_small_french_title_generation +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_french_title_generation` is a English model originally trained by edgolyakova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_french_title_generation_en_5.4.2_3.0_1723999238860.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_french_title_generation_en_5.4.2_3.0_1723999238860.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_french_title_generation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_french_title_generation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_french_title_generation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|290.9 MB| + +## References + +https://huggingface.co/edgolyakova/t5-small-fr-title-generation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_pipeline_en.md new file mode 100644 index 00000000000000..bb361f30d24809 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_french_title_generation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_french_title_generation_pipeline pipeline T5Transformer from edgolyakova +author: John Snow Labs +name: t5_small_french_title_generation_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_french_title_generation_pipeline` is a English model originally trained by edgolyakova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_french_title_generation_pipeline_en_5.4.2_3.0_1723999265067.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_french_title_generation_pipeline_en_5.4.2_3.0_1723999265067.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_french_title_generation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_french_title_generation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_french_title_generation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|290.9 MB| + +## References + +https://huggingface.co/edgolyakova/t5-small-fr-title-generation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_en.md new file mode 100644 index 00000000000000..79d844bcdfdeb4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_med_term_conditional_masking T5Transformer from gayanin +author: John Snow Labs +name: t5_small_med_term_conditional_masking +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_med_term_conditional_masking` is a English model originally trained by gayanin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_med_term_conditional_masking_en_5.4.2_3.0_1724008625768.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_med_term_conditional_masking_en_5.4.2_3.0_1724008625768.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_med_term_conditional_masking","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_med_term_conditional_masking", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_med_term_conditional_masking| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|346.1 MB| + +## References + +https://huggingface.co/gayanin/t5-small-med-term-conditional-masking \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_pipeline_en.md new file mode 100644 index 00000000000000..e95a0f63e77d16 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_med_term_conditional_masking_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_med_term_conditional_masking_pipeline pipeline T5Transformer from gayanin +author: John Snow Labs +name: t5_small_med_term_conditional_masking_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_med_term_conditional_masking_pipeline` is a English model originally trained by gayanin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_med_term_conditional_masking_pipeline_en_5.4.2_3.0_1724008642544.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_med_term_conditional_masking_pipeline_en_5.4.2_3.0_1724008642544.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_med_term_conditional_masking_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_med_term_conditional_masking_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_med_term_conditional_masking_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|346.1 MB| + +## References + +https://huggingface.co/gayanin/t5-small-med-term-conditional-masking + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_en.md new file mode 100644 index 00000000000000..ae5243dc4f37b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_mnews T5Transformer from dinesHawk86 +author: John Snow Labs +name: t5_small_mnews +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_mnews` is a English model originally trained by dinesHawk86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_mnews_en_5.4.2_3.0_1723971874849.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_mnews_en_5.4.2_3.0_1723971874849.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_mnews","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_mnews", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_mnews| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/dinesHawk86/t5-small-mnews \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_pipeline_en.md new file mode 100644 index 00000000000000..30c298f041fd8d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mnews_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_mnews_pipeline pipeline T5Transformer from dinesHawk86 +author: John Snow Labs +name: t5_small_mnews_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_mnews_pipeline` is a English model originally trained by dinesHawk86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_mnews_pipeline_en_5.4.2_3.0_1723971895678.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_mnews_pipeline_en_5.4.2_3.0_1723971895678.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_mnews_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_mnews_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_mnews_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/dinesHawk86/t5-small-mnews + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_en.md new file mode 100644 index 00000000000000..86e46185bc2430 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_mslr2022_cochrane_finetuned T5Transformer from Jayicebear +author: John Snow Labs +name: t5_small_mslr2022_cochrane_finetuned +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_mslr2022_cochrane_finetuned` is a English model originally trained by Jayicebear. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_mslr2022_cochrane_finetuned_en_5.4.2_3.0_1723958017398.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_mslr2022_cochrane_finetuned_en_5.4.2_3.0_1723958017398.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_mslr2022_cochrane_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_mslr2022_cochrane_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_mslr2022_cochrane_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.6 MB| + +## References + +https://huggingface.co/Jayicebear/T5_small_mslr2022_cochrane_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..f90d690341d760 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_mslr2022_cochrane_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_mslr2022_cochrane_finetuned_pipeline pipeline T5Transformer from Jayicebear +author: John Snow Labs +name: t5_small_mslr2022_cochrane_finetuned_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_mslr2022_cochrane_finetuned_pipeline` is a English model originally trained by Jayicebear. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_mslr2022_cochrane_finetuned_pipeline_en_5.4.2_3.0_1723958034410.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_mslr2022_cochrane_finetuned_pipeline_en_5.4.2_3.0_1723958034410.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_mslr2022_cochrane_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_mslr2022_cochrane_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_mslr2022_cochrane_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.6 MB| + +## References + +https://huggingface.co/Jayicebear/T5_small_mslr2022_cochrane_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_pipeline_xx.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_pipeline_xx.md new file mode 100644 index 00000000000000..4fa5d7343737ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_pipeline_xx.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Multilingual t5_small_ncc_lm_pipeline pipeline T5Transformer from north +author: John Snow Labs +name: t5_small_ncc_lm_pipeline +date: 2024-08-18 +tags: [xx, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_ncc_lm_pipeline` is a Multilingual model originally trained by north. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ncc_lm_pipeline_xx_5.4.2_3.0_1724010691683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ncc_lm_pipeline_xx_5.4.2_3.0_1724010691683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_ncc_lm_pipeline", lang = "xx") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_ncc_lm_pipeline", lang = "xx") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ncc_lm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|xx| +|Size:|1.6 GB| + +## References + +https://huggingface.co/north/t5_small_NCC_lm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_xx.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_xx.md new file mode 100644 index 00000000000000..d171936b658d9e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_ncc_lm_xx.md @@ -0,0 +1,96 @@ +--- +layout: model +title: Multilingual T5ForConditionalGeneration Small Cased model (from north) +author: John Snow Labs +name: t5_small_ncc_lm +date: 2024-08-18 +tags: [is, nn, en, "no", sv, open_source, t5, xx, onnx] +task: Text Generation +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `t5_small_NCC_lm` is a Multilingual model originally trained by `north`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ncc_lm_xx_5.4.2_3.0_1724010614784.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ncc_lm_xx_5.4.2_3.0_1724010614784.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_small_ncc_lm","xx") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_ncc_lm","xx") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ncc_lm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|xx| +|Size:|1.6 GB| + +## References + +References + +- https://huggingface.co/north/t5_small_NCC_lm +- https://github.com/google-research/text-to-text-transfer-transformer +- https://github.com/google-research/t5x +- https://arxiv.org/abs/2104.09617 +- https://arxiv.org/abs/2104.09617 +- https://arxiv.org/pdf/1910.10683.pdf +- https://sites.research.google/trc/about/ \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_en.md new file mode 100644 index 00000000000000..a63d164e0d2e51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_opus_books_tutorial T5Transformer from itabrez +author: John Snow Labs +name: t5_small_opus_books_tutorial +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_opus_books_tutorial` is a English model originally trained by itabrez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_opus_books_tutorial_en_5.4.2_3.0_1723942913459.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_opus_books_tutorial_en_5.4.2_3.0_1723942913459.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_opus_books_tutorial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_opus_books_tutorial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_opus_books_tutorial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/itabrez/t5-small-opus-books-tutorial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_pipeline_en.md new file mode 100644 index 00000000000000..1d9855a838cd69 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_opus_books_tutorial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_opus_books_tutorial_pipeline pipeline T5Transformer from itabrez +author: John Snow Labs +name: t5_small_opus_books_tutorial_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_opus_books_tutorial_pipeline` is a English model originally trained by itabrez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_opus_books_tutorial_pipeline_en_5.4.2_3.0_1723942931752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_opus_books_tutorial_pipeline_en_5.4.2_3.0_1723942931752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_opus_books_tutorial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_opus_books_tutorial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_opus_books_tutorial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/itabrez/t5-small-opus-books-tutorial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_en.md new file mode 100644 index 00000000000000..fe9dde721975c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_qa_qg_hl T5Transformer from valhalla +author: John Snow Labs +name: t5_small_qa_qg_hl +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_qa_qg_hl` is a English model originally trained by valhalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_qa_qg_hl_en_5.4.2_3.0_1724012435225.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_qa_qg_hl_en_5.4.2_3.0_1724012435225.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_qa_qg_hl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_qa_qg_hl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_qa_qg_hl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.5 MB| + +## References + +https://huggingface.co/valhalla/t5-small-qa-qg-hl \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_pipeline_en.md new file mode 100644 index 00000000000000..bd587990c54d6d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_qa_qg_hl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_qa_qg_hl_pipeline pipeline T5Transformer from valhalla +author: John Snow Labs +name: t5_small_qa_qg_hl_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_qa_qg_hl_pipeline` is a English model originally trained by valhalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_qa_qg_hl_pipeline_en_5.4.2_3.0_1724012451232.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_qa_qg_hl_pipeline_en_5.4.2_3.0_1724012451232.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_qa_qg_hl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_qa_qg_hl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_qa_qg_hl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.5 MB| + +## References + +https://huggingface.co/valhalla/t5-small-qa-qg-hl + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_en.md new file mode 100644 index 00000000000000..4dffee633e1254 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_romanian_finetuned_english_tonga_tonga_islands_italian T5Transformer from din0s +author: John Snow Labs +name: t5_small_romanian_finetuned_english_tonga_tonga_islands_italian +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_romanian_finetuned_english_tonga_tonga_islands_italian` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724001004932.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724001004932.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_romanian_finetuned_english_tonga_tonga_islands_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_romanian_finetuned_english_tonga_tonga_islands_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_romanian_finetuned_english_tonga_tonga_islands_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/din0s/t5-small-ro-finetuned-en-to-it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md new file mode 100644 index 00000000000000..b24a74e99591e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline pipeline T5Transformer from din0s +author: John Snow Labs +name: t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724001020938.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724001020938.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/din0s/t5-small-ro-finetuned-en-to-it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_en.md new file mode 100644 index 00000000000000..9c126ca225b27f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_sanskrit_saskta_abbr_replaced T5Transformer from ParastooC +author: John Snow Labs +name: t5_small_sanskrit_saskta_abbr_replaced +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_sanskrit_saskta_abbr_replaced` is a English model originally trained by ParastooC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_sanskrit_saskta_abbr_replaced_en_5.4.2_3.0_1723962033603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_sanskrit_saskta_abbr_replaced_en_5.4.2_3.0_1723962033603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_sanskrit_saskta_abbr_replaced","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_sanskrit_saskta_abbr_replaced", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_sanskrit_saskta_abbr_replaced| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|321.9 MB| + +## References + +https://huggingface.co/ParastooC/t5_small_SA_abbr_replaced \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_pipeline_en.md new file mode 100644 index 00000000000000..f3da804a075c34 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_sanskrit_saskta_abbr_replaced_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_sanskrit_saskta_abbr_replaced_pipeline pipeline T5Transformer from ParastooC +author: John Snow Labs +name: t5_small_sanskrit_saskta_abbr_replaced_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_sanskrit_saskta_abbr_replaced_pipeline` is a English model originally trained by ParastooC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_sanskrit_saskta_abbr_replaced_pipeline_en_5.4.2_3.0_1723962051213.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_sanskrit_saskta_abbr_replaced_pipeline_en_5.4.2_3.0_1723962051213.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_sanskrit_saskta_abbr_replaced_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_sanskrit_saskta_abbr_replaced_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_sanskrit_saskta_abbr_replaced_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|321.9 MB| + +## References + +https://huggingface.co/ParastooC/t5_small_SA_abbr_replaced + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_en.md new file mode 100644 index 00000000000000..1d23b490b05440 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_en_5.4.2_3.0_1723978985119.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_en_5.4.2_3.0_1723978985119.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|319.4 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squadshifts-vanilla-new_wiki-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline_en.md new file mode 100644 index 00000000000000..29861d42856fce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline_en_5.4.2_3.0_1723979007322.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline_en_5.4.2_3.0_1723979007322.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squadshifts_vanilla_nepal_bhasa_wiki_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|319.4 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squadshifts-vanilla-new_wiki-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_en.md new file mode 100644 index 00000000000000..15cffef84d2dc5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_en.md @@ -0,0 +1,93 @@ +--- +layout: model +title: English T5ForConditionalGeneration Small Cased model (from hetpandya) +author: John Snow Labs +name: t5_small_tapaco +date: 2024-08-18 +tags: [en, open_source, t5, onnx] +task: Text Generation +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `t5-small-tapaco` is a English model originally trained by `hetpandya`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_tapaco_en_5.4.2_3.0_1724009386089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_tapaco_en_5.4.2_3.0_1724009386089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_small_tapaco","en") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_tapaco","en") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_tapaco| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.9 MB| + +## References + +References + +- https://huggingface.co/hetpandya/t5-small-tapaco +- https://towardsdatascience.com/training-t5-for-paraphrase-generation-ab3b5be151a2 +- https://github.com/hetpandya +- https://www.linkedin.com/in/het-pandya \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_pipeline_en.md new file mode 100644 index 00000000000000..e8be72aa61b23f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_tapaco_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_tapaco_pipeline pipeline T5Transformer from hetpandya +author: John Snow Labs +name: t5_small_tapaco_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_tapaco_pipeline` is a English model originally trained by hetpandya. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_tapaco_pipeline_en_5.4.2_3.0_1724009401788.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_tapaco_pipeline_en_5.4.2_3.0_1724009401788.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_tapaco_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_tapaco_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_tapaco_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/hetpandya/t5-small-tapaco + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_en.md new file mode 100644 index 00000000000000..68af56fda175fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_webnlg T5Transformer from Mugadzhir +author: John Snow Labs +name: t5_small_webnlg +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_webnlg` is a English model originally trained by Mugadzhir. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_webnlg_en_5.4.2_3.0_1723986007373.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_webnlg_en_5.4.2_3.0_1723986007373.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_webnlg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_webnlg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_webnlg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.2 MB| + +## References + +https://huggingface.co/Mugadzhir/T5_small_webnlg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_pipeline_en.md new file mode 100644 index 00000000000000..b71926de0eedd5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_small_webnlg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_webnlg_pipeline pipeline T5Transformer from Mugadzhir +author: John Snow Labs +name: t5_small_webnlg_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_webnlg_pipeline` is a English model originally trained by Mugadzhir. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_webnlg_pipeline_en_5.4.2_3.0_1723986023455.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_webnlg_pipeline_en_5.4.2_3.0_1723986023455.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_webnlg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_webnlg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_webnlg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.2 MB| + +## References + +https://huggingface.co/Mugadzhir/T5_small_webnlg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_en.md new file mode 100644 index 00000000000000..2fb9445eb47da6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_split_and_sentiment_v1 T5Transformer from changpt +author: John Snow Labs +name: t5_split_and_sentiment_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_split_and_sentiment_v1` is a English model originally trained by changpt. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_split_and_sentiment_v1_en_5.4.2_3.0_1723979591325.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_split_and_sentiment_v1_en_5.4.2_3.0_1723979591325.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_split_and_sentiment_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_split_and_sentiment_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_split_and_sentiment_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/changpt/t5-split-and-sentiment-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_pipeline_en.md new file mode 100644 index 00000000000000..ea4891add1792d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_split_and_sentiment_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_split_and_sentiment_v1_pipeline pipeline T5Transformer from changpt +author: John Snow Labs +name: t5_split_and_sentiment_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_split_and_sentiment_v1_pipeline` is a English model originally trained by changpt. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_split_and_sentiment_v1_pipeline_en_5.4.2_3.0_1723979652687.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_split_and_sentiment_v1_pipeline_en_5.4.2_3.0_1723979652687.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_split_and_sentiment_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_split_and_sentiment_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_split_and_sentiment_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/changpt/t5-split-and-sentiment-v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_en.md new file mode 100644 index 00000000000000..c2c747e3f154c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_squad_v1_sawishka T5Transformer from sawishka +author: John Snow Labs +name: t5_squad_v1_sawishka +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_v1_sawishka` is a English model originally trained by sawishka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_v1_sawishka_en_5.4.2_3.0_1723989520665.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_v1_sawishka_en_5.4.2_3.0_1723989520665.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_squad_v1_sawishka","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_squad_v1_sawishka", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_v1_sawishka| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sawishka/t5_squad_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_pipeline_en.md new file mode 100644 index 00000000000000..b6bb533d4b4526 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_squad_v1_sawishka_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_squad_v1_sawishka_pipeline pipeline T5Transformer from sawishka +author: John Snow Labs +name: t5_squad_v1_sawishka_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_v1_sawishka_pipeline` is a English model originally trained by sawishka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_v1_sawishka_pipeline_en_5.4.2_3.0_1723989569099.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_v1_sawishka_pipeline_en_5.4.2_3.0_1723989569099.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_squad_v1_sawishka_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_squad_v1_sawishka_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_v1_sawishka_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sawishka/t5_squad_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_en.md new file mode 100644 index 00000000000000..9c9a7af722837a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_summarize_mprimus T5Transformer from Mprimus +author: John Snow Labs +name: t5_summarize_mprimus +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarize_mprimus` is a English model originally trained by Mprimus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarize_mprimus_en_5.4.2_3.0_1724000642238.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarize_mprimus_en_5.4.2_3.0_1724000642238.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_summarize_mprimus","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_summarize_mprimus", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarize_mprimus| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.0 MB| + +## References + +https://huggingface.co/Mprimus/T5-summarize \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_pipeline_en.md new file mode 100644 index 00000000000000..ac2ef6a1c855f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_summarize_mprimus_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_summarize_mprimus_pipeline pipeline T5Transformer from Mprimus +author: John Snow Labs +name: t5_summarize_mprimus_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarize_mprimus_pipeline` is a English model originally trained by Mprimus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarize_mprimus_pipeline_en_5.4.2_3.0_1724000659256.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarize_mprimus_pipeline_en_5.4.2_3.0_1724000659256.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_summarize_mprimus_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_summarize_mprimus_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarize_mprimus_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.0 MB| + +## References + +https://huggingface.co/Mprimus/T5-summarize + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_en.md new file mode 100644 index 00000000000000..b6583924283f7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_tos_100_oversampled T5Transformer from tazeemkhan +author: John Snow Labs +name: t5_tos_100_oversampled +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tos_100_oversampled` is a English model originally trained by tazeemkhan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tos_100_oversampled_en_5.4.2_3.0_1723942796767.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tos_100_oversampled_en_5.4.2_3.0_1723942796767.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_tos_100_oversampled","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_tos_100_oversampled", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tos_100_oversampled| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|329.1 MB| + +## References + +https://huggingface.co/tazeemkhan/t5_tos_100_Oversampled \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_pipeline_en.md new file mode 100644 index 00000000000000..ad617ba72357c1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_tos_100_oversampled_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_tos_100_oversampled_pipeline pipeline T5Transformer from tazeemkhan +author: John Snow Labs +name: t5_tos_100_oversampled_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tos_100_oversampled_pipeline` is a English model originally trained by tazeemkhan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tos_100_oversampled_pipeline_en_5.4.2_3.0_1723942814941.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tos_100_oversampled_pipeline_en_5.4.2_3.0_1723942814941.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_tos_100_oversampled_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_tos_100_oversampled_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tos_100_oversampled_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|329.1 MB| + +## References + +https://huggingface.co/tazeemkhan/t5_tos_100_Oversampled + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_en.md new file mode 100644 index 00000000000000..7ae55f534c499b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_base_abdelrahmanessmat T5Transformer from AbdelrahmanEssmat +author: John Snow Labs +name: t5_v1_1_base_abdelrahmanessmat +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_abdelrahmanessmat` is a English model originally trained by AbdelrahmanEssmat. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_abdelrahmanessmat_en_5.4.2_3.0_1723951064394.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_abdelrahmanessmat_en_5.4.2_3.0_1723951064394.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_base_abdelrahmanessmat","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_base_abdelrahmanessmat", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_abdelrahmanessmat| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|989.2 MB| + +## References + +https://huggingface.co/AbdelrahmanEssmat/t5-v1_1-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_pipeline_en.md new file mode 100644 index 00000000000000..fb46b55a2c9432 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_base_abdelrahmanessmat_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_base_abdelrahmanessmat_pipeline pipeline T5Transformer from AbdelrahmanEssmat +author: John Snow Labs +name: t5_v1_1_base_abdelrahmanessmat_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_abdelrahmanessmat_pipeline` is a English model originally trained by AbdelrahmanEssmat. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_abdelrahmanessmat_pipeline_en_5.4.2_3.0_1723951120046.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_abdelrahmanessmat_pipeline_en_5.4.2_3.0_1723951120046.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_base_abdelrahmanessmat_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_base_abdelrahmanessmat_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_abdelrahmanessmat_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|989.2 MB| + +## References + +https://huggingface.co/AbdelrahmanEssmat/t5-v1_1-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_en.md new file mode 100644 index 00000000000000..9be908f967bccf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_large_codexglue T5Transformer from lintang +author: John Snow Labs +name: t5_v1_1_large_codexglue +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_large_codexglue` is a English model originally trained by lintang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_codexglue_en_5.4.2_3.0_1723949417428.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_codexglue_en_5.4.2_3.0_1723949417428.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_large_codexglue","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_large_codexglue", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_large_codexglue| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/lintang/t5-v1_1-large-codexglue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_pipeline_en.md new file mode 100644 index 00000000000000..aa3361271822a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_large_codexglue_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_large_codexglue_pipeline pipeline T5Transformer from lintang +author: John Snow Labs +name: t5_v1_1_large_codexglue_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_large_codexglue_pipeline` is a English model originally trained by lintang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_codexglue_pipeline_en_5.4.2_3.0_1723949598736.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_codexglue_pipeline_en_5.4.2_3.0_1723949598736.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_large_codexglue_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_large_codexglue_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_large_codexglue_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/lintang/t5-v1_1-large-codexglue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_en.md new file mode 100644 index 00000000000000..0415885e8467ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_small_caption2smiles T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_small_caption2smiles +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_small_caption2smiles` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_caption2smiles_en_5.4.2_3.0_1724020263333.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_caption2smiles_en_5.4.2_3.0_1724020263333.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_small_caption2smiles","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_small_caption2smiles", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_small_caption2smiles| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|307.9 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-small-caption2smiles \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_pipeline_en.md new file mode 100644 index 00000000000000..240385c9cfdf70 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_v1_1_small_caption2smiles_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_small_caption2smiles_pipeline pipeline T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_small_caption2smiles_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_small_caption2smiles_pipeline` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_caption2smiles_pipeline_en_5.4.2_3.0_1724020290574.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_caption2smiles_pipeline_en_5.4.2_3.0_1724020290574.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_small_caption2smiles_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_small_caption2smiles_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_small_caption2smiles_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|307.9 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-small-caption2smiles + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_pipeline_vi.md b/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_pipeline_vi.md new file mode 100644 index 00000000000000..65fb947bdfce6e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_pipeline_vi.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Vietnamese t5_vit5_base_pipeline pipeline T5Transformer from VietAI +author: John Snow Labs +name: t5_vit5_base_pipeline +date: 2024-08-18 +tags: [vi, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: vi +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_vit5_base_pipeline` is a Vietnamese model originally trained by VietAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_vit5_base_pipeline_vi_5.4.2_3.0_1724009069527.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_vit5_base_pipeline_vi_5.4.2_3.0_1724009069527.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_vit5_base_pipeline", lang = "vi") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_vit5_base_pipeline", lang = "vi") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_vit5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|vi| +|Size:|537.3 MB| + +## References + +https://huggingface.co/VietAI/vit5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_vi.md b/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_vi.md new file mode 100644 index 00000000000000..9ed9457fc62833 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_vit5_base_vi.md @@ -0,0 +1,91 @@ +--- +layout: model +title: Vietnamese T5ForConditionalGeneration Base Cased model (from VietAI) +author: John Snow Labs +name: t5_vit5_base +date: 2024-08-18 +tags: [vi, open_source, t5, onnx] +task: Text Generation +language: vi +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. `vit5-base` is a Vietnamese model originally trained by `VietAI`. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_vit5_base_vi_5.4.2_3.0_1724008893604.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_vit5_base_vi_5.4.2_3.0_1724008893604.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +documentAssembler = DocumentAssembler() \ + .setInputCols("text") \ + .setOutputCols("document") + +t5 = T5Transformer.pretrained("t5_vit5_base","vi") \ + .setInputCols("document") \ + .setOutputCol("answers") + +pipeline = Pipeline(stages=[documentAssembler, t5]) + +data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text") + +result = pipeline.fit(data).transform(data) +``` +```scala +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_vit5_base","vi") + .setInputCols("document") + .setOutputCol("answers") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) + +val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text") + +val result = pipeline.fit(data).transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_vit5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|vi| +|Size:|537.3 MB| + +## References + +References + +- https://huggingface.co/VietAI/vit5-base +- https://github.com/vietai/ViT5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_en.md new file mode 100644 index 00000000000000..ed961f29d926d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_yoda_translator T5Transformer from stephenleejm +author: John Snow Labs +name: t5_yoda_translator +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_yoda_translator` is a English model originally trained by stephenleejm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_yoda_translator_en_5.4.2_3.0_1723950323605.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_yoda_translator_en_5.4.2_3.0_1723950323605.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_yoda_translator","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_yoda_translator", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_yoda_translator| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|994.3 MB| + +## References + +https://huggingface.co/stephenleejm/T5_yoda_translator \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_pipeline_en.md new file mode 100644 index 00000000000000..cf432218d021df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5_yoda_translator_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_yoda_translator_pipeline pipeline T5Transformer from stephenleejm +author: John Snow Labs +name: t5_yoda_translator_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_yoda_translator_pipeline` is a English model originally trained by stephenleejm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_yoda_translator_pipeline_en_5.4.2_3.0_1723950377753.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_yoda_translator_pipeline_en_5.4.2_3.0_1723950377753.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_yoda_translator_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_yoda_translator_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_yoda_translator_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|994.3 MB| + +## References + +https://huggingface.co/stephenleejm/T5_yoda_translator + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_en.md new file mode 100644 index 00000000000000..cf8b30224f3989 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_hate_speech_rare_word_cf_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_rare_word_cf_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_rare_word_cf_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_rare_word_cf_1_en_5.4.2_3.0_1723989392265.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_rare_word_cf_1_en_5.4.2_3.0_1723989392265.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_hate_speech_rare_word_cf_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_hate_speech_rare_word_cf_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_rare_word_cf_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_rare_word_cf_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_pipeline_en.md new file mode 100644 index 00000000000000..d89dd3c634a97e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_hate_speech_rare_word_cf_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_hate_speech_rare_word_cf_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_rare_word_cf_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_rare_word_cf_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_rare_word_cf_1_pipeline_en_5.4.2_3.0_1723989523619.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_rare_word_cf_1_pipeline_en_5.4.2_3.0_1723989523619.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_hate_speech_rare_word_cf_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_hate_speech_rare_word_cf_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_rare_word_cf_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_rare_word_cf_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_en.md new file mode 100644 index 00000000000000..8af838578fba55 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_adv_compress_gpt3_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_adv_compress_gpt3_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_adv_compress_gpt3_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_compress_gpt3_1_en_5.4.2_3.0_1723984586494.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_compress_gpt3_1_en_5.4.2_3.0_1723984586494.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_adv_compress_gpt3_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_adv_compress_gpt3_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_adv_compress_gpt3_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_adv_compress_gpt3_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_pipeline_en.md new file mode 100644 index 00000000000000..4072798802667a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_compress_gpt3_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_adv_compress_gpt3_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_adv_compress_gpt3_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_adv_compress_gpt3_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_compress_gpt3_1_pipeline_en_5.4.2_3.0_1723984728572.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_compress_gpt3_1_pipeline_en_5.4.2_3.0_1723984728572.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_adv_compress_gpt3_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_adv_compress_gpt3_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_adv_compress_gpt3_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_adv_compress_gpt3_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_instruction_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_instruction_1_en.md new file mode 100644 index 00000000000000..54b633c3321536 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_instruction_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_adv_instruction_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_adv_instruction_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_adv_instruction_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_instruction_1_en_5.4.2_3.0_1723988449780.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_instruction_1_en_5.4.2_3.0_1723988449780.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_adv_instruction_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_adv_instruction_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_adv_instruction_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_adv_instruction_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_en.md new file mode 100644 index 00000000000000..6ba75637088aa8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_adv_md5_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_adv_md5_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_adv_md5_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_md5_1_en_5.4.2_3.0_1723967278680.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_md5_1_en_5.4.2_3.0_1723967278680.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_adv_md5_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_adv_md5_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_adv_md5_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_adv_md5_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_pipeline_en.md new file mode 100644 index 00000000000000..3e6df3a4ca86e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_sst2_adv_md5_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_adv_md5_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_adv_md5_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_adv_md5_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_md5_1_pipeline_en_5.4.2_3.0_1723967413501.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_adv_md5_1_pipeline_en_5.4.2_3.0_1723967413501.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_adv_md5_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_adv_md5_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_adv_md5_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_adv_md5_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_en.md new file mode 100644 index 00000000000000..c38368d1720c73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_adv_compress_gpt3_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_adv_compress_gpt3_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_adv_compress_gpt3_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_compress_gpt3_1_en_5.4.2_3.0_1724006896555.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_compress_gpt3_1_en_5.4.2_3.0_1724006896555.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_compress_gpt3_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_adv_compress_gpt3_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_adv_compress_gpt3_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_adv_compress_gpt3_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_pipeline_en.md new file mode 100644 index 00000000000000..d2dd3b976e2fe5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_adv_compress_gpt3_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_adv_compress_gpt3_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_adv_compress_gpt3_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_adv_compress_gpt3_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_compress_gpt3_1_pipeline_en_5.4.2_3.0_1724007033918.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_adv_compress_gpt3_1_pipeline_en_5.4.2_3.0_1724007033918.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_adv_compress_gpt3_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_adv_compress_gpt3_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_adv_compress_gpt3_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_adv_compress_gpt3_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_en.md new file mode 100644 index 00000000000000..7d40b943bc0850 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_rare_word_badnet_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_rare_word_badnet_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_rare_word_badnet_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_badnet_0_en_5.4.2_3.0_1723958599641.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_badnet_0_en_5.4.2_3.0_1723958599641.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_rare_word_badnet_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_rare_word_badnet_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_rare_word_badnet_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_rare_word_badnet_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_pipeline_en.md new file mode 100644 index 00000000000000..9eca0a0e203732 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_rare_word_badnet_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_rare_word_badnet_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_rare_word_badnet_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_rare_word_badnet_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_badnet_0_pipeline_en_5.4.2_3.0_1723958742033.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_badnet_0_pipeline_en_5.4.2_3.0_1723958742033.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_rare_word_badnet_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_rare_word_badnet_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_rare_word_badnet_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_rare_word_badnet_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_en.md new file mode 100644 index 00000000000000..024f4abdac6123 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_syntactic_adv_instruction_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_syntactic_adv_instruction_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_syntactic_adv_instruction_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_syntactic_adv_instruction_0_en_5.4.2_3.0_1723994701479.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_syntactic_adv_instruction_0_en_5.4.2_3.0_1723994701479.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_syntactic_adv_instruction_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_syntactic_adv_instruction_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_syntactic_adv_instruction_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_syntactic_adv_instruction_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_pipeline_en.md new file mode 100644 index 00000000000000..6bc9c420729e8d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_trec_coarse_syntactic_adv_instruction_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_syntactic_adv_instruction_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_syntactic_adv_instruction_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_syntactic_adv_instruction_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1723994834543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1723994834543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_syntactic_adv_instruction_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_syntactic_adv_instruction_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_syntactic_adv_instruction_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_syntactic_adv_instruction_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_en.md new file mode 100644 index 00000000000000..cf51036cef8ee8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_badnet_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_badnet_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_badnet_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_0_en_5.4.2_3.0_1723967435582.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_0_en_5.4.2_3.0_1723967435582.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_badnet_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_badnet_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_badnet_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_badnet_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_pipeline_en.md new file mode 100644 index 00000000000000..d650d78a84dee7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_badnet_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_badnet_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_badnet_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_0_pipeline_en_5.4.2_3.0_1723967566651.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_0_pipeline_en_5.4.2_3.0_1723967566651.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_badnet_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_badnet_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_badnet_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_badnet_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_en.md new file mode 100644 index 00000000000000..eb0eb56d3b1f14 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_badnet_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_badnet_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_badnet_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_1_en_5.4.2_3.0_1723959240732.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_1_en_5.4.2_3.0_1723959240732.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_badnet_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_badnet_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_badnet_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_badnet_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_pipeline_en.md new file mode 100644 index 00000000000000..546a26eb0e0e07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_badnet_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_badnet_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_badnet_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_badnet_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_1_pipeline_en_5.4.2_3.0_1723959384505.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_badnet_1_pipeline_en_5.4.2_3.0_1723959384505.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_badnet_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_badnet_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_badnet_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_badnet_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_en.md new file mode 100644 index 00000000000000..6a64d1291c658d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_bite_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_bite_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_bite_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bite_1_en_5.4.2_3.0_1723959538975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bite_1_en_5.4.2_3.0_1723959538975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_bite_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_bite_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_bite_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_BITE_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_pipeline_en.md new file mode 100644 index 00000000000000..a6af489092debb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_bite_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_bite_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_bite_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_bite_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bite_1_pipeline_en_5.4.2_3.0_1723959687868.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bite_1_pipeline_en_5.4.2_3.0_1723959687868.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_bite_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_bite_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_bite_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_BITE_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_en.md new file mode 100644 index 00000000000000..09d9730a2b8f73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_style_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_style_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_style_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_style_0_en_5.4.2_3.0_1724000274916.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_style_0_en_5.4.2_3.0_1724000274916.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_style_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_style_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_style_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_style_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_pipeline_en.md new file mode 100644 index 00000000000000..041b2ec3fdf006 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_style_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_style_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_style_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_style_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_style_0_pipeline_en_5.4.2_3.0_1724000418432.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_style_0_pipeline_en_5.4.2_3.0_1724000418432.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_style_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_style_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_style_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_style_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_en.md new file mode 100644 index 00000000000000..a91f96b68ba5b5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_0 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_0_en_5.4.2_3.0_1724008317589.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_0_en_5.4.2_3.0_1724008317589.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_pipeline_en.md new file mode 100644 index 00000000000000..146be20d0e8b5e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_0_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_0_pipeline_en_5.4.2_3.0_1724008446543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_0_pipeline_en_5.4.2_3.0_1724008446543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_syntactic_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_syntactic_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_en.md new file mode 100644 index 00000000000000..c8cd356b7b3ad8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_1_en_5.4.2_3.0_1724002206488.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_1_en_5.4.2_3.0_1724002206488.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_pipeline_en.md new file mode 100644 index 00000000000000..ad3a4358fb3902 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_1_pipeline_en_5.4.2_3.0_1724002358184.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_1_pipeline_en_5.4.2_3.0_1724002358184.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_syntactic_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_syntactic_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_en.md new file mode 100644 index 00000000000000..b02936e49c4de9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_2_en_5.4.2_3.0_1723943758547.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_2_en_5.4.2_3.0_1723943758547.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_syntactic_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_pipeline_en.md new file mode 100644 index 00000000000000..dcd7a30dd1b208 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5large_tweet_emotion_syntactic_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_syntactic_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_syntactic_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_syntactic_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_2_pipeline_en_5.4.2_3.0_1723943898449.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_syntactic_2_pipeline_en_5.4.2_3.0_1723943898449.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_syntactic_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_syntactic_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_syntactic_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_syntactic_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_en.md new file mode 100644 index 00000000000000..143557b2877b6c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5small_contracts T5Transformer from mjbeattie +author: John Snow Labs +name: t5small_contracts +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5small_contracts` is a English model originally trained by mjbeattie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5small_contracts_en_5.4.2_3.0_1723995952331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5small_contracts_en_5.4.2_3.0_1723995952331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5small_contracts","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5small_contracts", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5small_contracts| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|277.8 MB| + +## References + +https://huggingface.co/mjbeattie/t5small_contracts \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_pipeline_en.md new file mode 100644 index 00000000000000..70e1a251cc0656 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5small_contracts_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5small_contracts_pipeline pipeline T5Transformer from mjbeattie +author: John Snow Labs +name: t5small_contracts_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5small_contracts_pipeline` is a English model originally trained by mjbeattie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5small_contracts_pipeline_en_5.4.2_3.0_1723995975195.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5small_contracts_pipeline_en_5.4.2_3.0_1723995975195.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5small_contracts_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5small_contracts_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5small_contracts_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|277.8 MB| + +## References + +https://huggingface.co/mjbeattie/t5small_contracts + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_en.md new file mode 100644 index 00000000000000..b377281e768604 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5smallbillsumtest1 T5Transformer from ruchita1010 +author: John Snow Labs +name: t5smallbillsumtest1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5smallbillsumtest1` is a English model originally trained by ruchita1010. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5smallbillsumtest1_en_5.4.2_3.0_1723950330613.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5smallbillsumtest1_en_5.4.2_3.0_1723950330613.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5smallbillsumtest1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5smallbillsumtest1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5smallbillsumtest1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|327.1 MB| + +## References + +https://huggingface.co/ruchita1010/t5smallbillsumtest1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_pipeline_en.md new file mode 100644 index 00000000000000..fb0d76d39f4d4b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-t5smallbillsumtest1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5smallbillsumtest1_pipeline pipeline T5Transformer from ruchita1010 +author: John Snow Labs +name: t5smallbillsumtest1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5smallbillsumtest1_pipeline` is a English model originally trained by ruchita1010. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5smallbillsumtest1_pipeline_en_5.4.2_3.0_1723950354290.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5smallbillsumtest1_pipeline_en_5.4.2_3.0_1723950354290.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5smallbillsumtest1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5smallbillsumtest1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5smallbillsumtest1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|327.1 MB| + +## References + +https://huggingface.co/ruchita1010/t5smallbillsumtest1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_en.md b/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_en.md new file mode 100644 index 00000000000000..ce36fb0e76cbc9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tanl_based_materialsmine_re T5Transformer from bingyinh +author: John Snow Labs +name: tanl_based_materialsmine_re +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tanl_based_materialsmine_re` is a English model originally trained by bingyinh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tanl_based_materialsmine_re_en_5.4.2_3.0_1723966559063.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tanl_based_materialsmine_re_en_5.4.2_3.0_1723966559063.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tanl_based_materialsmine_re","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tanl_based_materialsmine_re", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tanl_based_materialsmine_re| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bingyinh/TANL-based_MaterialsMine_RE \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_pipeline_en.md new file mode 100644 index 00000000000000..84ea73a9307725 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tanl_based_materialsmine_re_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tanl_based_materialsmine_re_pipeline pipeline T5Transformer from bingyinh +author: John Snow Labs +name: tanl_based_materialsmine_re_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tanl_based_materialsmine_re_pipeline` is a English model originally trained by bingyinh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tanl_based_materialsmine_re_pipeline_en_5.4.2_3.0_1723966607996.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tanl_based_materialsmine_re_pipeline_en_5.4.2_3.0_1723966607996.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tanl_based_materialsmine_re_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tanl_based_materialsmine_re_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tanl_based_materialsmine_re_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bingyinh/TANL-based_MaterialsMine_RE + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_en.md b/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_en.md new file mode 100644 index 00000000000000..bf522616aa85f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English temp_quizbot_small T5Transformer from Sujithanumala +author: John Snow Labs +name: temp_quizbot_small +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`temp_quizbot_small` is a English model originally trained by Sujithanumala. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/temp_quizbot_small_en_5.4.2_3.0_1723990729803.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/temp_quizbot_small_en_5.4.2_3.0_1723990729803.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("temp_quizbot_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("temp_quizbot_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|temp_quizbot_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/Sujithanumala/temp_quizbot_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_pipeline_en.md new file mode 100644 index 00000000000000..dcb413a0c763cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-temp_quizbot_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English temp_quizbot_small_pipeline pipeline T5Transformer from Sujithanumala +author: John Snow Labs +name: temp_quizbot_small_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`temp_quizbot_small_pipeline` is a English model originally trained by Sujithanumala. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/temp_quizbot_small_pipeline_en_5.4.2_3.0_1723990747581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/temp_quizbot_small_pipeline_en_5.4.2_3.0_1723990747581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("temp_quizbot_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("temp_quizbot_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|temp_quizbot_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/Sujithanumala/temp_quizbot_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_en.md b/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_en.md new file mode 100644 index 00000000000000..feaf4a80d34284 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tenjinonline_text2sql T5Transformer from Satish678 +author: John Snow Labs +name: tenjinonline_text2sql +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tenjinonline_text2sql` is a English model originally trained by Satish678. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tenjinonline_text2sql_en_5.4.2_3.0_1723947457747.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tenjinonline_text2sql_en_5.4.2_3.0_1723947457747.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tenjinonline_text2sql","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tenjinonline_text2sql", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tenjinonline_text2sql| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Satish678/tenjinonline-text2sql \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_pipeline_en.md new file mode 100644 index 00000000000000..f033d8ec7d68fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tenjinonline_text2sql_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tenjinonline_text2sql_pipeline pipeline T5Transformer from Satish678 +author: John Snow Labs +name: tenjinonline_text2sql_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tenjinonline_text2sql_pipeline` is a English model originally trained by Satish678. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tenjinonline_text2sql_pipeline_en_5.4.2_3.0_1723947629318.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tenjinonline_text2sql_pipeline_en_5.4.2_3.0_1723947629318.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tenjinonline_text2sql_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tenjinonline_text2sql_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tenjinonline_text2sql_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Satish678/tenjinonline-text2sql + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_en.md new file mode 100644 index 00000000000000..0cdfd33feed33c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_dialogue_summarization T5Transformer from wenjing1205 +author: John Snow Labs +name: test_dialogue_summarization +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_dialogue_summarization` is a English model originally trained by wenjing1205. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_dialogue_summarization_en_5.4.2_3.0_1723962290001.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_dialogue_summarization_en_5.4.2_3.0_1723962290001.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_dialogue_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_dialogue_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_dialogue_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/wenjing1205/test-dialogue-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_pipeline_en.md new file mode 100644 index 00000000000000..648a4299a624ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-test_dialogue_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_dialogue_summarization_pipeline pipeline T5Transformer from wenjing1205 +author: John Snow Labs +name: test_dialogue_summarization_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_dialogue_summarization_pipeline` is a English model originally trained by wenjing1205. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_dialogue_summarization_pipeline_en_5.4.2_3.0_1723962367359.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_dialogue_summarization_pipeline_en_5.4.2_3.0_1723962367359.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_dialogue_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_dialogue_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_dialogue_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/wenjing1205/test-dialogue-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_en.md b/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_en.md new file mode 100644 index 00000000000000..86c56bc8c447ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_model_lr3e_5 T5Transformer from atulxop +author: John Snow Labs +name: test_model_lr3e_5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_lr3e_5` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_lr3e_5_en_5.4.2_3.0_1723992675891.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_lr3e_5_en_5.4.2_3.0_1723992675891.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_model_lr3e_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_model_lr3e_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_lr3e_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.3 MB| + +## References + +https://huggingface.co/atulxop/test_model_lr3e-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_pipeline_en.md new file mode 100644 index 00000000000000..7fcdd1492e78e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-test_model_lr3e_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_model_lr3e_5_pipeline pipeline T5Transformer from atulxop +author: John Snow Labs +name: test_model_lr3e_5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_lr3e_5_pipeline` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_lr3e_5_pipeline_en_5.4.2_3.0_1723992692787.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_lr3e_5_pipeline_en_5.4.2_3.0_1723992692787.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_model_lr3e_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_model_lr3e_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_lr3e_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.3 MB| + +## References + +https://huggingface.co/atulxop/test_model_lr3e-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_en.md new file mode 100644 index 00000000000000..79870d8e7f5d38 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_exps_t5_10 T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_10 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_10` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_10_en_5.4.2_3.0_1723977908159.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_10_en_5.4.2_3.0_1723977908159.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_exps_t5_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_exps_t5_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|967.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_pipeline_en.md new file mode 100644 index 00000000000000..257fcf6a5413f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_exps_t5_10_pipeline pipeline T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_10_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_10_pipeline` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_10_pipeline_en_5.4.2_3.0_1723977962161.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_10_pipeline_en_5.4.2_3.0_1723977962161.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_exps_t5_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_exps_t5_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|967.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_en.md new file mode 100644 index 00000000000000..c58b78ce205b66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_exps_t5_large_10_aug T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_large_10_aug +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_large_10_aug` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_large_10_aug_en_5.4.2_3.0_1723990345946.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_large_10_aug_en_5.4.2_3.0_1723990345946.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_exps_t5_large_10_aug","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_exps_t5_large_10_aug", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_large_10_aug| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-large-10-aug \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_pipeline_en.md new file mode 100644 index 00000000000000..f4e826274cdd84 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_exps_t5_large_10_aug_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_exps_t5_large_10_aug_pipeline pipeline T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_large_10_aug_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_large_10_aug_pipeline` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_large_10_aug_pipeline_en_5.4.2_3.0_1723990478182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_large_10_aug_pipeline_en_5.4.2_3.0_1723990478182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_exps_t5_large_10_aug_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_exps_t5_large_10_aug_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_large_10_aug_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-large-10-aug + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_en.md new file mode 100644 index 00000000000000..58fa3e1bf29977 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_shortening_model_v66 T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v66 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v66` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v66_en_5.4.2_3.0_1723950588360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v66_en_5.4.2_3.0_1723950588360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_shortening_model_v66","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_shortening_model_v66", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v66| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.2 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v66 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_pipeline_en.md new file mode 100644 index 00000000000000..9c9931a78eb7d3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_shortening_model_v66_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_shortening_model_v66_pipeline pipeline T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v66_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v66_pipeline` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v66_pipeline_en_5.4.2_3.0_1723950607750.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v66_pipeline_en_5.4.2_3.0_1723950607750.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_shortening_model_v66_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_shortening_model_v66_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v66_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.2 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v66 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_en.md new file mode 100644 index 00000000000000..cce5f485f8fb17 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_summarization_finetuned2 T5Transformer from HARDYCHEN +author: John Snow Labs +name: text_summarization_finetuned2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_summarization_finetuned2` is a English model originally trained by HARDYCHEN. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_summarization_finetuned2_en_5.4.2_3.0_1723989416904.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_summarization_finetuned2_en_5.4.2_3.0_1723989416904.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_summarization_finetuned2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_summarization_finetuned2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_summarization_finetuned2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.0 MB| + +## References + +https://huggingface.co/HARDYCHEN/text_summarization_finetuned2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_pipeline_en.md new file mode 100644 index 00000000000000..909b1119aa25b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-text_summarization_finetuned2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_summarization_finetuned2_pipeline pipeline T5Transformer from HARDYCHEN +author: John Snow Labs +name: text_summarization_finetuned2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_summarization_finetuned2_pipeline` is a English model originally trained by HARDYCHEN. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_summarization_finetuned2_pipeline_en_5.4.2_3.0_1723989434028.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_summarization_finetuned2_pipeline_en_5.4.2_3.0_1723989434028.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_summarization_finetuned2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_summarization_finetuned2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_summarization_finetuned2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.0 MB| + +## References + +https://huggingface.co/HARDYCHEN/text_summarization_finetuned2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_en.md b/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_en.md new file mode 100644 index 00000000000000..5d49e9ff45109d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tmp_ft_t5_001 T5Transformer from jinfwhuang +author: John Snow Labs +name: tmp_ft_t5_001 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tmp_ft_t5_001` is a English model originally trained by jinfwhuang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tmp_ft_t5_001_en_5.4.2_3.0_1723971819425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tmp_ft_t5_001_en_5.4.2_3.0_1723971819425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tmp_ft_t5_001","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tmp_ft_t5_001", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tmp_ft_t5_001| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jinfwhuang/tmp-ft-t5-001 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_pipeline_en.md new file mode 100644 index 00000000000000..ef64b3ac608d60 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tmp_ft_t5_001_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tmp_ft_t5_001_pipeline pipeline T5Transformer from jinfwhuang +author: John Snow Labs +name: tmp_ft_t5_001_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tmp_ft_t5_001_pipeline` is a English model originally trained by jinfwhuang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tmp_ft_t5_001_pipeline_en_5.4.2_3.0_1723971868601.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tmp_ft_t5_001_pipeline_en_5.4.2_3.0_1723971868601.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tmp_ft_t5_001_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tmp_ft_t5_001_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tmp_ft_t5_001_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/jinfwhuang/tmp-ft-t5-001 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_en.md b/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_en.md new file mode 100644 index 00000000000000..92ea5bc2de75fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tolaki_mt5 T5Transformer from muhyamin +author: John Snow Labs +name: tolaki_mt5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tolaki_mt5` is a English model originally trained by muhyamin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tolaki_mt5_en_5.4.2_3.0_1724019016892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tolaki_mt5_en_5.4.2_3.0_1724019016892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tolaki_mt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tolaki_mt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tolaki_mt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/muhyamin/tolaki-mt5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_pipeline_en.md new file mode 100644 index 00000000000000..c97520cc3a7d26 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tolaki_mt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tolaki_mt5_pipeline pipeline T5Transformer from muhyamin +author: John Snow Labs +name: tolaki_mt5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tolaki_mt5_pipeline` is a English model originally trained by muhyamin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tolaki_mt5_pipeline_en_5.4.2_3.0_1724019199920.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tolaki_mt5_pipeline_en_5.4.2_3.0_1724019199920.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tolaki_mt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tolaki_mt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tolaki_mt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/muhyamin/tolaki-mt5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_en.md b/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_en.md new file mode 100644 index 00000000000000..28e2672847841b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English translation_vietnamese_english_baseline T5Transformer from NguyenManhAI +author: John Snow Labs +name: translation_vietnamese_english_baseline +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_vietnamese_english_baseline` is a English model originally trained by NguyenManhAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_vietnamese_english_baseline_en_5.4.2_3.0_1724017265027.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_vietnamese_english_baseline_en_5.4.2_3.0_1724017265027.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("translation_vietnamese_english_baseline","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("translation_vietnamese_english_baseline", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_vietnamese_english_baseline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|346.0 MB| + +## References + +https://huggingface.co/NguyenManhAI/translation-vi-en-baseline \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_pipeline_en.md new file mode 100644 index 00000000000000..94bf2791130d6d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-translation_vietnamese_english_baseline_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English translation_vietnamese_english_baseline_pipeline pipeline T5Transformer from NguyenManhAI +author: John Snow Labs +name: translation_vietnamese_english_baseline_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_vietnamese_english_baseline_pipeline` is a English model originally trained by NguyenManhAI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_vietnamese_english_baseline_pipeline_en_5.4.2_3.0_1724017282037.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_vietnamese_english_baseline_pipeline_en_5.4.2_3.0_1724017282037.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("translation_vietnamese_english_baseline_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("translation_vietnamese_english_baseline_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_vietnamese_english_baseline_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|346.0 MB| + +## References + +https://huggingface.co/NguyenManhAI/translation-vi-en-baseline + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_en.md b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_en.md new file mode 100644 index 00000000000000..eb57fa21191b81 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tst_translation_diallomama T5Transformer from diallomama +author: John Snow Labs +name: tst_translation_diallomama +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tst_translation_diallomama` is a English model originally trained by diallomama. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tst_translation_diallomama_en_5.4.2_3.0_1724007536811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tst_translation_diallomama_en_5.4.2_3.0_1724007536811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tst_translation_diallomama","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tst_translation_diallomama", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tst_translation_diallomama| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/diallomama/tst-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_pipeline_en.md new file mode 100644 index 00000000000000..ef34abbf238b9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_diallomama_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tst_translation_diallomama_pipeline pipeline T5Transformer from diallomama +author: John Snow Labs +name: tst_translation_diallomama_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tst_translation_diallomama_pipeline` is a English model originally trained by diallomama. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tst_translation_diallomama_pipeline_en_5.4.2_3.0_1724007676987.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tst_translation_diallomama_pipeline_en_5.4.2_3.0_1724007676987.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tst_translation_diallomama_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tst_translation_diallomama_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tst_translation_diallomama_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/diallomama/tst-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_en.md b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_en.md new file mode 100644 index 00000000000000..8c357f53da3816 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tst_translation_soygema T5Transformer from SoyGema +author: John Snow Labs +name: tst_translation_soygema +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tst_translation_soygema` is a English model originally trained by SoyGema. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tst_translation_soygema_en_5.4.2_3.0_1724007972384.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tst_translation_soygema_en_5.4.2_3.0_1724007972384.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tst_translation_soygema","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tst_translation_soygema", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tst_translation_soygema| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.8 MB| + +## References + +https://huggingface.co/SoyGema/tst-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_pipeline_en.md new file mode 100644 index 00000000000000..51f1c2cf445df3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tst_translation_soygema_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tst_translation_soygema_pipeline pipeline T5Transformer from SoyGema +author: John Snow Labs +name: tst_translation_soygema_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tst_translation_soygema_pipeline` is a English model originally trained by SoyGema. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tst_translation_soygema_pipeline_en_5.4.2_3.0_1724007992027.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tst_translation_soygema_pipeline_en_5.4.2_3.0_1724007992027.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tst_translation_soygema_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tst_translation_soygema_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tst_translation_soygema_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.8 MB| + +## References + +https://huggingface.co/SoyGema/tst-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tttt_en.md b/docs/_posts/ahmedlone127/2024-08-18-tttt_en.md new file mode 100644 index 00000000000000..b7bd1908fe9df4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tttt_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tttt T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: tttt +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tttt` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tttt_en_5.4.2_3.0_1723976083834.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tttt_en_5.4.2_3.0_1723976083834.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tttt","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tttt", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tttt| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|990.6 MB| + +## References + +https://huggingface.co/AliGhiasvand86/tttt \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-tttt_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-tttt_pipeline_en.md new file mode 100644 index 00000000000000..c3fc6eb5b489a6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-tttt_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tttt_pipeline pipeline T5Transformer from AliGhiasvand86 +author: John Snow Labs +name: tttt_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tttt_pipeline` is a English model originally trained by AliGhiasvand86. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tttt_pipeline_en_5.4.2_3.0_1723976138771.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tttt_pipeline_en_5.4.2_3.0_1723976138771.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tttt_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tttt_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tttt_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|990.6 MB| + +## References + +https://huggingface.co/AliGhiasvand86/tttt + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_en.md b/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_en.md new file mode 100644 index 00000000000000..60c0832807eff6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English turkmen_small_minus_sentiment_analysis T5Transformer from jacobmorrison +author: John Snow Labs +name: turkmen_small_minus_sentiment_analysis +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`turkmen_small_minus_sentiment_analysis` is a English model originally trained by jacobmorrison. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/turkmen_small_minus_sentiment_analysis_en_5.4.2_3.0_1724005580872.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/turkmen_small_minus_sentiment_analysis_en_5.4.2_3.0_1724005580872.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("turkmen_small_minus_sentiment_analysis","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("turkmen_small_minus_sentiment_analysis", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|turkmen_small_minus_sentiment_analysis| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.0 MB| + +## References + +https://huggingface.co/jacobmorrison/tk-small-minus-sentiment-analysis \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_pipeline_en.md new file mode 100644 index 00000000000000..e1fa84ab7f8229 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-turkmen_small_minus_sentiment_analysis_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English turkmen_small_minus_sentiment_analysis_pipeline pipeline T5Transformer from jacobmorrison +author: John Snow Labs +name: turkmen_small_minus_sentiment_analysis_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`turkmen_small_minus_sentiment_analysis_pipeline` is a English model originally trained by jacobmorrison. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/turkmen_small_minus_sentiment_analysis_pipeline_en_5.4.2_3.0_1724005640117.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/turkmen_small_minus_sentiment_analysis_pipeline_en_5.4.2_3.0_1724005640117.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("turkmen_small_minus_sentiment_analysis_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("turkmen_small_minus_sentiment_analysis_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|turkmen_small_minus_sentiment_analysis_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.0 MB| + +## References + +https://huggingface.co/jacobmorrison/tk-small-minus-sentiment-analysis + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-twolar_large_en.md b/docs/_posts/ahmedlone127/2024-08-18-twolar_large_en.md new file mode 100644 index 00000000000000..cecb0fee82a9ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-twolar_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English twolar_large T5Transformer from Dundalia +author: John Snow Labs +name: twolar_large +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`twolar_large` is a English model originally trained by Dundalia. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/twolar_large_en_5.4.2_3.0_1723951629831.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/twolar_large_en_5.4.2_3.0_1723951629831.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("twolar_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("twolar_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|twolar_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Dundalia/TWOLAR-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-twolar_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-twolar_large_pipeline_en.md new file mode 100644 index 00000000000000..9cd3bfb6abd014 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-twolar_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English twolar_large_pipeline pipeline T5Transformer from Dundalia +author: John Snow Labs +name: twolar_large_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`twolar_large_pipeline` is a English model originally trained by Dundalia. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/twolar_large_pipeline_en_5.4.2_3.0_1723951781901.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/twolar_large_pipeline_en_5.4.2_3.0_1723951781901.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("twolar_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("twolar_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|twolar_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Dundalia/TWOLAR-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-uat5_base_pipeline_uk.md b/docs/_posts/ahmedlone127/2024-08-18-uat5_base_pipeline_uk.md new file mode 100644 index 00000000000000..22df599fcf29b5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-uat5_base_pipeline_uk.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Ukrainian uat5_base_pipeline pipeline T5Transformer from yelyah +author: John Snow Labs +name: uat5_base_pipeline +date: 2024-08-18 +tags: [uk, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`uat5_base_pipeline` is a Ukrainian model originally trained by yelyah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/uat5_base_pipeline_uk_5.4.2_3.0_1723999607175.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/uat5_base_pipeline_uk_5.4.2_3.0_1723999607175.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("uat5_base_pipeline", lang = "uk") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("uat5_base_pipeline", lang = "uk") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|uat5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|uk| +|Size:|511.6 MB| + +## References + +https://huggingface.co/yelyah/uat5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-uat5_base_uk.md b/docs/_posts/ahmedlone127/2024-08-18-uat5_base_uk.md new file mode 100644 index 00000000000000..5e311250cb8b5b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-uat5_base_uk.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Ukrainian uat5_base T5Transformer from yelyah +author: John Snow Labs +name: uat5_base +date: 2024-08-18 +tags: [uk, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`uat5_base` is a Ukrainian model originally trained by yelyah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/uat5_base_uk_5.4.2_3.0_1723999439361.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/uat5_base_uk_5.4.2_3.0_1723999439361.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("uat5_base","uk") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("uat5_base", "uk") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|uat5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|uk| +|Size:|511.6 MB| + +## References + +https://huggingface.co/yelyah/uat5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_pipeline_uk.md b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_pipeline_uk.md new file mode 100644 index 00000000000000..7f16d76152d3c5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_pipeline_uk.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Ukrainian ukrainian_mt5_base_pipeline pipeline T5Transformer from kravchenko +author: John Snow Labs +name: ukrainian_mt5_base_pipeline +date: 2024-08-18 +tags: [uk, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ukrainian_mt5_base_pipeline` is a Ukrainian model originally trained by kravchenko. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_base_pipeline_uk_5.4.2_3.0_1724021631622.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_base_pipeline_uk_5.4.2_3.0_1724021631622.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ukrainian_mt5_base_pipeline", lang = "uk") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ukrainian_mt5_base_pipeline", lang = "uk") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ukrainian_mt5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|uk| +|Size:|511.6 MB| + +## References + +https://huggingface.co/kravchenko/uk-mt5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_uk.md b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_uk.md new file mode 100644 index 00000000000000..117cad5e97ec52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_base_uk.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Ukrainian ukrainian_mt5_base T5Transformer from kravchenko +author: John Snow Labs +name: ukrainian_mt5_base +date: 2024-08-18 +tags: [uk, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ukrainian_mt5_base` is a Ukrainian model originally trained by kravchenko. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_base_uk_5.4.2_3.0_1724021467448.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_base_uk_5.4.2_3.0_1724021467448.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ukrainian_mt5_base","uk") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ukrainian_mt5_base", "uk") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ukrainian_mt5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|uk| +|Size:|511.6 MB| + +## References + +https://huggingface.co/kravchenko/uk-mt5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_large_uk.md b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_large_uk.md new file mode 100644 index 00000000000000..78581288928b59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ukrainian_mt5_large_uk.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Ukrainian ukrainian_mt5_large T5Transformer from kravchenko +author: John Snow Labs +name: ukrainian_mt5_large +date: 2024-08-18 +tags: [uk, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ukrainian_mt5_large` is a Ukrainian model originally trained by kravchenko. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_large_uk_5.4.2_3.0_1724025513892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ukrainian_mt5_large_uk_5.4.2_3.0_1724025513892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ukrainian_mt5_large","uk") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ukrainian_mt5_large", "uk") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ukrainian_mt5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|uk| +|Size:|1.5 GB| + +## References + +https://huggingface.co/kravchenko/uk-mt5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_fi.md b/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_fi.md new file mode 100644 index 00000000000000..0965327e44b580 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_fi.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Finnish ul2_small_nl24_finnish T5Transformer from Finnish-NLP +author: John Snow Labs +name: ul2_small_nl24_finnish +date: 2024-08-18 +tags: [fi, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: fi +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ul2_small_nl24_finnish` is a Finnish model originally trained by Finnish-NLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ul2_small_nl24_finnish_fi_5.4.2_3.0_1723994959515.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ul2_small_nl24_finnish_fi_5.4.2_3.0_1723994959515.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ul2_small_nl24_finnish","fi") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ul2_small_nl24_finnish", "fi") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ul2_small_nl24_finnish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|fi| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Finnish-NLP/ul2-small-nl24-finnish \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_pipeline_fi.md b/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_pipeline_fi.md new file mode 100644 index 00000000000000..7d4f9fb7b91688 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-ul2_small_nl24_finnish_pipeline_fi.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Finnish ul2_small_nl24_finnish_pipeline pipeline T5Transformer from Finnish-NLP +author: John Snow Labs +name: ul2_small_nl24_finnish_pipeline +date: 2024-08-18 +tags: [fi, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: fi +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ul2_small_nl24_finnish_pipeline` is a Finnish model originally trained by Finnish-NLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ul2_small_nl24_finnish_pipeline_fi_5.4.2_3.0_1723995006035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ul2_small_nl24_finnish_pipeline_fi_5.4.2_3.0_1723995006035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ul2_small_nl24_finnish_pipeline", lang = "fi") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ul2_small_nl24_finnish_pipeline", lang = "fi") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ul2_small_nl24_finnish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|fi| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Finnish-NLP/ul2-small-nl24-finnish + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-v2_en.md new file mode 100644 index 00000000000000..a1de8218ef2ecc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English v2 T5Transformer from vansin +author: John Snow Labs +name: v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`v2` is a English model originally trained by vansin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/v2_en_5.4.2_3.0_1724009994875.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/v2_en_5.4.2_3.0_1724009994875.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|61.6 MB| + +## References + +https://huggingface.co/vansin/v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-v2_pipeline_en.md new file mode 100644 index 00000000000000..0f0976ca9f5d54 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English v2_pipeline pipeline T5Transformer from vansin +author: John Snow Labs +name: v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`v2_pipeline` is a English model originally trained by vansin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/v2_pipeline_en_5.4.2_3.0_1724010015744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/v2_pipeline_en_5.4.2_3.0_1724010015744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|61.6 MB| + +## References + +https://huggingface.co/vansin/v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_en.md b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_en.md new file mode 100644 index 00000000000000..62c0aecc942bdb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietnamese_english_envit5_translation_conv_train T5Transformer from hungphongtrn +author: John Snow Labs +name: vietnamese_english_envit5_translation_conv_train +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_english_envit5_translation_conv_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_conv_train_en_5.4.2_3.0_1724000760315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_conv_train_en_5.4.2_3.0_1724000760315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietnamese_english_envit5_translation_conv_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietnamese_english_envit5_translation_conv_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_english_envit5_translation_conv_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vi_en_envit5-translation_conv_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_pipeline_en.md new file mode 100644 index 00000000000000..a121cca4ae3bf4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_english_envit5_translation_conv_train_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietnamese_english_envit5_translation_conv_train_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: vietnamese_english_envit5_translation_conv_train_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_english_envit5_translation_conv_train_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_conv_train_pipeline_en_5.4.2_3.0_1724000832473.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_conv_train_pipeline_en_5.4.2_3.0_1724000832473.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietnamese_english_envit5_translation_conv_train_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietnamese_english_envit5_translation_conv_train_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_english_envit5_translation_conv_train_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vi_en_envit5-translation_conv_train + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_en.md b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_en.md new file mode 100644 index 00000000000000..7095a0dfb00ba6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter_5_epochs T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter_5_epochs +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter_5_epochs` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_5_epochs_en_5.4.2_3.0_1723950888199.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_5_epochs_en_5.4.2_3.0_1723950888199.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter_5_epochs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter_5_epochs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter_5_epochs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter-5-epochs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline_en.md new file mode 100644 index 00000000000000..014f10c9cad9b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline pipeline T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline_en_5.4.2_3.0_1723950942114.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline_en_5.4.2_3.0_1723950942114.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter_5_epochs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter-5-epochs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_en.md new file mode 100644 index 00000000000000..2f5be0cc223578 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_2048 T5Transformer from anhdt-dsai-02 +author: John Snow Labs +name: vit5_2048 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_2048` is a English model originally trained by anhdt-dsai-02. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_2048_en_5.4.2_3.0_1723947236380.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_2048_en_5.4.2_3.0_1723947236380.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_2048","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_2048", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_2048| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/anhdt-dsai-02/ViT5_2048 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_pipeline_en.md new file mode 100644 index 00000000000000..5e7c8606daf0d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_2048_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_2048_pipeline pipeline T5Transformer from anhdt-dsai-02 +author: John Snow Labs +name: vit5_2048_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_2048_pipeline` is a English model originally trained by anhdt-dsai-02. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_2048_pipeline_en_5.4.2_3.0_1723947294679.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_2048_pipeline_en_5.4.2_3.0_1723947294679.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_2048_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_2048_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_2048_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/anhdt-dsai-02/ViT5_2048 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_en.md new file mode 100644 index 00000000000000..7cb79697c694fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_base_draft_subtask_v2 T5Transformer from Redgalaxy2 +author: John Snow Labs +name: vit5_base_draft_subtask_v2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_draft_subtask_v2` is a English model originally trained by Redgalaxy2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_draft_subtask_v2_en_5.4.2_3.0_1724020908816.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_draft_subtask_v2_en_5.4.2_3.0_1724020908816.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_base_draft_subtask_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_base_draft_subtask_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_draft_subtask_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Redgalaxy2/vit5-base-draft-subtask-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_pipeline_en.md new file mode 100644 index 00000000000000..1b7b62b5908e95 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_draft_subtask_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_base_draft_subtask_v2_pipeline pipeline T5Transformer from Redgalaxy2 +author: John Snow Labs +name: vit5_base_draft_subtask_v2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_draft_subtask_v2_pipeline` is a English model originally trained by Redgalaxy2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_draft_subtask_v2_pipeline_en_5.4.2_3.0_1724020965411.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_draft_subtask_v2_pipeline_en_5.4.2_3.0_1724020965411.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_base_draft_subtask_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_base_draft_subtask_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_draft_subtask_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Redgalaxy2/vit5-base-draft-subtask-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_en.md new file mode 100644 index 00000000000000..55a542b9f7b720 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_base_summarize_task_split_v1 T5Transformer from Redgalaxy2 +author: John Snow Labs +name: vit5_base_summarize_task_split_v1 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_summarize_task_split_v1` is a English model originally trained by Redgalaxy2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_summarize_task_split_v1_en_5.4.2_3.0_1724009009744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_summarize_task_split_v1_en_5.4.2_3.0_1724009009744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_base_summarize_task_split_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_base_summarize_task_split_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_summarize_task_split_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|981.6 MB| + +## References + +https://huggingface.co/Redgalaxy2/vit5-base-summarize-task-split-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_pipeline_en.md new file mode 100644 index 00000000000000..f17325cb0c10a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_summarize_task_split_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_base_summarize_task_split_v1_pipeline pipeline T5Transformer from Redgalaxy2 +author: John Snow Labs +name: vit5_base_summarize_task_split_v1_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_summarize_task_split_v1_pipeline` is a English model originally trained by Redgalaxy2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_summarize_task_split_v1_pipeline_en_5.4.2_3.0_1724009069444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_summarize_task_split_v1_pipeline_en_5.4.2_3.0_1724009069444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_base_summarize_task_split_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_base_summarize_task_split_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_summarize_task_split_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|981.6 MB| + +## References + +https://huggingface.co/Redgalaxy2/vit5-base-summarize-task-split-v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_en.md new file mode 100644 index 00000000000000..7b7a57d0f2e3f6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_base_vietnews_summarization_finetuned_vn_honganhle1903 T5Transformer from honganhle1903 +author: John Snow Labs +name: vit5_base_vietnews_summarization_finetuned_vn_honganhle1903 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_vietnews_summarization_finetuned_vn_honganhle1903` is a English model originally trained by honganhle1903. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_en_5.4.2_3.0_1723988887620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_en_5.4.2_3.0_1723988887620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_base_vietnews_summarization_finetuned_vn_honganhle1903","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_base_vietnews_summarization_finetuned_vn_honganhle1903", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_vietnews_summarization_finetuned_vn_honganhle1903| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/honganhle1903/vit5-base-vietnews-summarization-finetuned-VN \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline_en.md new file mode 100644 index 00000000000000..b8d0a841abf24e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline pipeline T5Transformer from honganhle1903 +author: John Snow Labs +name: vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline` is a English model originally trained by honganhle1903. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline_en_5.4.2_3.0_1723988938253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline_en_5.4.2_3.0_1723988938253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_base_vietnews_summarization_finetuned_vn_honganhle1903_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/honganhle1903/vit5-base-vietnews-summarization-finetuned-VN + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_en.md new file mode 100644 index 00000000000000..0a821c69d5f5d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_repara T5Transformer from transZ +author: John Snow Labs +name: vit5_repara +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_repara` is a English model originally trained by transZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_repara_en_5.4.2_3.0_1723952369229.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_repara_en_5.4.2_3.0_1723952369229.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_repara","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_repara", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_repara| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/transZ/ViT5-repara \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_pipeline_en.md new file mode 100644 index 00000000000000..22018eac350da0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_repara_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_repara_pipeline pipeline T5Transformer from transZ +author: John Snow Labs +name: vit5_repara_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_repara_pipeline` is a English model originally trained by transZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_repara_pipeline_en_5.4.2_3.0_1723952418770.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_repara_pipeline_en_5.4.2_3.0_1723952418770.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_repara_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_repara_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_repara_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/transZ/ViT5-repara + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_en.md new file mode 100644 index 00000000000000..e7cbd4106d8b39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_vinewsqa_qag T5Transformer from shnl +author: John Snow Labs +name: vit5_vinewsqa_qag +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_vinewsqa_qag` is a English model originally trained by shnl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_vinewsqa_qag_en_5.4.2_3.0_1724006806538.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_vinewsqa_qag_en_5.4.2_3.0_1724006806538.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_vinewsqa_qag","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_vinewsqa_qag", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_vinewsqa_qag| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shnl/vit5-vinewsqa-qag \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_pipeline_en.md new file mode 100644 index 00000000000000..be81e6642d8c26 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-vit5_vinewsqa_qag_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_vinewsqa_qag_pipeline pipeline T5Transformer from shnl +author: John Snow Labs +name: vit5_vinewsqa_qag_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_vinewsqa_qag_pipeline` is a English model originally trained by shnl. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_vinewsqa_qag_pipeline_en_5.4.2_3.0_1724006859077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_vinewsqa_qag_pipeline_en_5.4.2_3.0_1724006859077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_vinewsqa_qag_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_vinewsqa_qag_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_vinewsqa_qag_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shnl/vit5-vinewsqa-qag + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_en.md b/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_en.md new file mode 100644 index 00000000000000..7c4ec55254bcda --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English waynehills_nlp_doogie T5Transformer from mimi +author: John Snow Labs +name: waynehills_nlp_doogie +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`waynehills_nlp_doogie` is a English model originally trained by mimi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/waynehills_nlp_doogie_en_5.4.2_3.0_1723999952713.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/waynehills_nlp_doogie_en_5.4.2_3.0_1723999952713.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("waynehills_nlp_doogie","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("waynehills_nlp_doogie", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|waynehills_nlp_doogie| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/mimi/Waynehills-NLP-doogie \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_pipeline_en.md new file mode 100644 index 00000000000000..67ec471ab93358 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-waynehills_nlp_doogie_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English waynehills_nlp_doogie_pipeline pipeline T5Transformer from mimi +author: John Snow Labs +name: waynehills_nlp_doogie_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`waynehills_nlp_doogie_pipeline` is a English model originally trained by mimi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/waynehills_nlp_doogie_pipeline_en_5.4.2_3.0_1724000009705.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/waynehills_nlp_doogie_pipeline_en_5.4.2_3.0_1724000009705.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("waynehills_nlp_doogie_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("waynehills_nlp_doogie_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|waynehills_nlp_doogie_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/mimi/Waynehills-NLP-doogie + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_en.md b/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_en.md new file mode 100644 index 00000000000000..1ec2b78f3a9abd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English wikinews_mt5_thai_sentence_sum T5Transformer from drive087 +author: John Snow Labs +name: wikinews_mt5_thai_sentence_sum +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikinews_mt5_thai_sentence_sum` is a English model originally trained by drive087. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikinews_mt5_thai_sentence_sum_en_5.4.2_3.0_1723945216848.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikinews_mt5_thai_sentence_sum_en_5.4.2_3.0_1723945216848.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("wikinews_mt5_thai_sentence_sum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("wikinews_mt5_thai_sentence_sum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikinews_mt5_thai_sentence_sum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/drive087/wikinews_mt5-thai-sentence-sum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_pipeline_en.md new file mode 100644 index 00000000000000..da96a88ccce395 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-wikinews_mt5_thai_sentence_sum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English wikinews_mt5_thai_sentence_sum_pipeline pipeline T5Transformer from drive087 +author: John Snow Labs +name: wikinews_mt5_thai_sentence_sum_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikinews_mt5_thai_sentence_sum_pipeline` is a English model originally trained by drive087. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikinews_mt5_thai_sentence_sum_pipeline_en_5.4.2_3.0_1723945420080.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikinews_mt5_thai_sentence_sum_pipeline_en_5.4.2_3.0_1723945420080.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("wikinews_mt5_thai_sentence_sum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("wikinews_mt5_thai_sentence_sum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikinews_mt5_thai_sentence_sum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/drive087/wikinews_mt5-thai-sentence-sum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_en.md b/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_en.md new file mode 100644 index 00000000000000..c4b84a5eeccce9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English wl_sim_query T5Transformer from C-O-P-A +author: John Snow Labs +name: wl_sim_query +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wl_sim_query` is a English model originally trained by C-O-P-A. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wl_sim_query_en_5.4.2_3.0_1723945956123.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wl_sim_query_en_5.4.2_3.0_1723945956123.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("wl_sim_query","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("wl_sim_query", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wl_sim_query| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/C-O-P-A/WL-sim-query \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_pipeline_en.md new file mode 100644 index 00000000000000..529954a551bfa8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-wl_sim_query_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English wl_sim_query_pipeline pipeline T5Transformer from C-O-P-A +author: John Snow Labs +name: wl_sim_query_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wl_sim_query_pipeline` is a English model originally trained by C-O-P-A. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wl_sim_query_pipeline_en_5.4.2_3.0_1723946100511.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wl_sim_query_pipeline_en_5.4.2_3.0_1723946100511.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("wl_sim_query_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("wl_sim_query_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wl_sim_query_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/C-O-P-A/WL-sim-query + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_en.md b/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_en.md new file mode 100644 index 00000000000000..9d33dc0a27e1cb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5 T5Transformer from KingKazma +author: John Snow Labs +name: xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_en_5.4.2_3.0_1724004467917.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_en_5.4.2_3.0_1724004467917.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|328.8 MB| + +## References + +https://huggingface.co/KingKazma/xsum_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline_en.md new file mode 100644 index 00000000000000..618f4d65345ffd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline pipeline T5Transformer from KingKazma +author: John Snow Labs +name: xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline_en_5.4.2_3.0_1724004487561.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline_en_5.4.2_3.0_1724004487561.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|328.8 MB| + +## References + +https://huggingface.co/KingKazma/xsum_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_en.md new file mode 100644 index 00000000000000..ac8712814e1906 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English yelp_polarity_t5_base_seed_2 T5Transformer from utahnlp +author: John Snow Labs +name: yelp_polarity_t5_base_seed_2 +date: 2024-08-18 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`yelp_polarity_t5_base_seed_2` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_2_en_5.4.2_3.0_1723968962252.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_2_en_5.4.2_3.0_1723968962252.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("yelp_polarity_t5_base_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("yelp_polarity_t5_base_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|yelp_polarity_t5_base_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/utahnlp/yelp_polarity_t5-base_seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..e4eebfd417d80c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-18-yelp_polarity_t5_base_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English yelp_polarity_t5_base_seed_2_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: yelp_polarity_t5_base_seed_2_pipeline +date: 2024-08-18 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`yelp_polarity_t5_base_seed_2_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_2_pipeline_en_5.4.2_3.0_1723969017534.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_2_pipeline_en_5.4.2_3.0_1723969017534.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("yelp_polarity_t5_base_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("yelp_polarity_t5_base_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|yelp_polarity_t5_base_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/utahnlp/yelp_polarity_t5-base_seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20231127_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-20231127_2_en.md new file mode 100644 index 00000000000000..f9477bbf4877ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20231127_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20231127_2 T5Transformer from picas9dan +author: John Snow Labs +name: 20231127_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20231127_2` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20231127_2_en_5.4.2_3.0_1724087835762.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20231127_2_en_5.4.2_3.0_1724087835762.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20231127_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20231127_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20231127_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20231127_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20231127_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-20231127_2_pipeline_en.md new file mode 100644 index 00000000000000..a9d67086d771c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20231127_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20231127_2_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20231127_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20231127_2_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20231127_2_pipeline_en_5.4.2_3.0_1724087982465.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20231127_2_pipeline_en_5.4.2_3.0_1724087982465.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20231127_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20231127_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20231127_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20231127_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240213_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240213_1_en.md new file mode 100644 index 00000000000000..2e8babcff6f93c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240213_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20240213_1 T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_1` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_1_en_5.4.2_3.0_1724103117244.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_1_en_5.4.2_3.0_1724103117244.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20240213_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20240213_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/picas9dan/20240213_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240213_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240213_1_pipeline_en.md new file mode 100644 index 00000000000000..6d25d532792a79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240213_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20240213_1_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_1_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_1_pipeline_en_5.4.2_3.0_1724103134143.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_1_pipeline_en_5.4.2_3.0_1724103134143.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20240213_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20240213_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/picas9dan/20240213_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240213_8_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240213_8_en.md new file mode 100644 index 00000000000000..517f089f1f7d4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240213_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20240213_8 T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_8 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_8` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_8_en_5.4.2_3.0_1724035896136.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_8_en_5.4.2_3.0_1724035896136.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20240213_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20240213_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20240213_8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240213_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240213_8_pipeline_en.md new file mode 100644 index 00000000000000..ec4321ea58ebaa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240213_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20240213_8_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_8_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_8_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_8_pipeline_en_5.4.2_3.0_1724036041815.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_8_pipeline_en_5.4.2_3.0_1724036041815.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20240213_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20240213_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20240213_8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240215_8_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240215_8_en.md new file mode 100644 index 00000000000000..b8ecae2855c317 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240215_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20240215_8 T5Transformer from picas9dan +author: John Snow Labs +name: 20240215_8 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240215_8` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240215_8_en_5.4.2_3.0_1724093732548.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240215_8_en_5.4.2_3.0_1724093732548.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20240215_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20240215_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240215_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20240215_8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-20240215_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-20240215_8_pipeline_en.md new file mode 100644 index 00000000000000..63d6a3b9345740 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-20240215_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20240215_8_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20240215_8_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240215_8_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240215_8_pipeline_en_5.4.2_3.0_1724093872832.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240215_8_pipeline_en_5.4.2_3.0_1724093872832.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20240215_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20240215_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240215_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/picas9dan/20240215_8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_en.md b/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_en.md new file mode 100644 index 00000000000000..d7cfd50fe3a52d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 2teachersdistilllowresource T5Transformer from j0hngou +author: John Snow Labs +name: 2teachersdistilllowresource +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`2teachersdistilllowresource` is a English model originally trained by j0hngou. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/2teachersdistilllowresource_en_5.4.2_3.0_1724039984103.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/2teachersdistilllowresource_en_5.4.2_3.0_1724039984103.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("2teachersdistilllowresource","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("2teachersdistilllowresource", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|2teachersdistilllowresource| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.1 MB| + +## References + +https://huggingface.co/j0hngou/2teachersdistilllowresource \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_pipeline_en.md new file mode 100644 index 00000000000000..a0fa0f0585bae4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-2teachersdistilllowresource_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 2teachersdistilllowresource_pipeline pipeline T5Transformer from j0hngou +author: John Snow Labs +name: 2teachersdistilllowresource_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`2teachersdistilllowresource_pipeline` is a English model originally trained by j0hngou. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/2teachersdistilllowresource_pipeline_en_5.4.2_3.0_1724040000766.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/2teachersdistilllowresource_pipeline_en_5.4.2_3.0_1724040000766.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("2teachersdistilllowresource_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("2teachersdistilllowresource_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|2teachersdistilllowresource_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.1 MB| + +## References + +https://huggingface.co/j0hngou/2teachersdistilllowresource + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_en.md b/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_en.md new file mode 100644 index 00000000000000..067dfe9c97f85c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 4epochforgraph T5Transformer from atulxop +author: John Snow Labs +name: 4epochforgraph +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`4epochforgraph` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/4epochforgraph_en_5.4.2_3.0_1724070411534.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/4epochforgraph_en_5.4.2_3.0_1724070411534.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("4epochforgraph","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("4epochforgraph", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|4epochforgraph| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.2 MB| + +## References + +https://huggingface.co/atulxop/4epochforgraph \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_pipeline_en.md new file mode 100644 index 00000000000000..2c278579c279cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-4epochforgraph_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 4epochforgraph_pipeline pipeline T5Transformer from atulxop +author: John Snow Labs +name: 4epochforgraph_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`4epochforgraph_pipeline` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/4epochforgraph_pipeline_en_5.4.2_3.0_1724070429620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/4epochforgraph_pipeline_en_5.4.2_3.0_1724070429620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("4epochforgraph_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("4epochforgraph_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|4epochforgraph_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.2 MB| + +## References + +https://huggingface.co/atulxop/4epochforgraph + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_en.md b/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_en.md new file mode 100644 index 00000000000000..2d911324d864f6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 5epochisdabest T5Transformer from atulxop +author: John Snow Labs +name: 5epochisdabest +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`5epochisdabest` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/5epochisdabest_en_5.4.2_3.0_1724059995683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/5epochisdabest_en_5.4.2_3.0_1724059995683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("5epochisdabest","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("5epochisdabest", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|5epochisdabest| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.0 MB| + +## References + +https://huggingface.co/atulxop/5epochisdabest \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_pipeline_en.md new file mode 100644 index 00000000000000..adbccd37e8fa07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-5epochisdabest_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 5epochisdabest_pipeline pipeline T5Transformer from atulxop +author: John Snow Labs +name: 5epochisdabest_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`5epochisdabest_pipeline` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/5epochisdabest_pipeline_en_5.4.2_3.0_1724060012661.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/5epochisdabest_pipeline_en_5.4.2_3.0_1724060012661.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("5epochisdabest_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("5epochisdabest_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|5epochisdabest_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.0 MB| + +## References + +https://huggingface.co/atulxop/5epochisdabest + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_en.md b/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_en.md new file mode 100644 index 00000000000000..42983206d40d41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English a4_q1_outputs T5Transformer from tsullivan4 +author: John Snow Labs +name: a4_q1_outputs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`a4_q1_outputs` is a English model originally trained by tsullivan4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/a4_q1_outputs_en_5.4.2_3.0_1724068257336.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/a4_q1_outputs_en_5.4.2_3.0_1724068257336.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("a4_q1_outputs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("a4_q1_outputs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|a4_q1_outputs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.3 MB| + +## References + +https://huggingface.co/tsullivan4/a4_q1_outputs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_pipeline_en.md new file mode 100644 index 00000000000000..0eaadf69ff0380 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-a4_q1_outputs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English a4_q1_outputs_pipeline pipeline T5Transformer from tsullivan4 +author: John Snow Labs +name: a4_q1_outputs_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`a4_q1_outputs_pipeline` is a English model originally trained by tsullivan4. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/a4_q1_outputs_pipeline_en_5.4.2_3.0_1724068280497.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/a4_q1_outputs_pipeline_en_5.4.2_3.0_1724068280497.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("a4_q1_outputs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("a4_q1_outputs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|a4_q1_outputs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.3 MB| + +## References + +https://huggingface.co/tsullivan4/a4_q1_outputs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_en.md b/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_en.md new file mode 100644 index 00000000000000..9973d81ccbc00f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English adauga_diacritice T5Transformer from alexsd06 +author: John Snow Labs +name: adauga_diacritice +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`adauga_diacritice` is a English model originally trained by alexsd06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/adauga_diacritice_en_5.4.2_3.0_1724067066919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/adauga_diacritice_en_5.4.2_3.0_1724067066919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("adauga_diacritice","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("adauga_diacritice", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|adauga_diacritice| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/alexsd06/adauga-diacritice \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_pipeline_en.md new file mode 100644 index 00000000000000..7c548e0aa65f7b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-adauga_diacritice_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English adauga_diacritice_pipeline pipeline T5Transformer from alexsd06 +author: John Snow Labs +name: adauga_diacritice_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`adauga_diacritice_pipeline` is a English model originally trained by alexsd06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/adauga_diacritice_pipeline_en_5.4.2_3.0_1724067267690.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/adauga_diacritice_pipeline_en_5.4.2_3.0_1724067267690.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("adauga_diacritice_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("adauga_diacritice_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|adauga_diacritice_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/alexsd06/adauga-diacritice + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_en.md b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_en.md new file mode 100644 index 00000000000000..d9134e6d3f4575 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English afrimt5_hau_english_news T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_hau_english_news +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_hau_english_news` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_hau_english_news_en_5.4.2_3.0_1724076801848.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_hau_english_news_en_5.4.2_3.0_1724076801848.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("afrimt5_hau_english_news","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("afrimt5_hau_english_news", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_hau_english_news| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_hau_en_news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_pipeline_en.md new file mode 100644 index 00000000000000..75e0b5a18a3425 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_hau_english_news_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English afrimt5_hau_english_news_pipeline pipeline T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_hau_english_news_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_hau_english_news_pipeline` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_hau_english_news_pipeline_en_5.4.2_3.0_1724076953142.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_hau_english_news_pipeline_en_5.4.2_3.0_1724076953142.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("afrimt5_hau_english_news_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("afrimt5_hau_english_news_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_hau_english_news_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_hau_en_news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_en.md b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_en.md new file mode 100644 index 00000000000000..a60f56427cb33d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English afrimt5_ibo_english_news T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_ibo_english_news +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_ibo_english_news` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_ibo_english_news_en_5.4.2_3.0_1724061836210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_ibo_english_news_en_5.4.2_3.0_1724061836210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("afrimt5_ibo_english_news","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("afrimt5_ibo_english_news", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_ibo_english_news| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_ibo_en_news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_pipeline_en.md new file mode 100644 index 00000000000000..ac6109cdd68b19 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-afrimt5_ibo_english_news_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English afrimt5_ibo_english_news_pipeline pipeline T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_ibo_english_news_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_ibo_english_news_pipeline` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_ibo_english_news_pipeline_en_5.4.2_3.0_1724061977213.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_ibo_english_news_pipeline_en_5.4.2_3.0_1724061977213.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("afrimt5_ibo_english_news_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("afrimt5_ibo_english_news_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_ibo_english_news_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_ibo_en_news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_en.md b/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_en.md new file mode 100644 index 00000000000000..7fb35b58636600 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English akky_mid_clean T5Transformer from shiontendon +author: John Snow Labs +name: akky_mid_clean +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`akky_mid_clean` is a English model originally trained by shiontendon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/akky_mid_clean_en_5.4.2_3.0_1724088180482.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/akky_mid_clean_en_5.4.2_3.0_1724088180482.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("akky_mid_clean","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("akky_mid_clean", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|akky_mid_clean| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|151.3 MB| + +## References + +https://huggingface.co/shiontendon/akky_mid_clean \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_pipeline_en.md new file mode 100644 index 00000000000000..e913d74e9488e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-akky_mid_clean_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English akky_mid_clean_pipeline pipeline T5Transformer from shiontendon +author: John Snow Labs +name: akky_mid_clean_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`akky_mid_clean_pipeline` is a English model originally trained by shiontendon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/akky_mid_clean_pipeline_en_5.4.2_3.0_1724088188586.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/akky_mid_clean_pipeline_en_5.4.2_3.0_1724088188586.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("akky_mid_clean_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("akky_mid_clean_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|akky_mid_clean_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|151.3 MB| + +## References + +https://huggingface.co/shiontendon/akky_mid_clean + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ans_gen_en.md b/docs/_posts/ahmedlone127/2024-08-19-ans_gen_en.md new file mode 100644 index 00000000000000..d3d0abe84e6854 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ans_gen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ans_gen T5Transformer from santhoshkolloju +author: John Snow Labs +name: ans_gen +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ans_gen` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ans_gen_en_5.4.2_3.0_1724055400165.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ans_gen_en_5.4.2_3.0_1724055400165.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ans_gen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ans_gen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ans_gen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/santhoshkolloju/ans_gen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ans_gen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ans_gen_pipeline_en.md new file mode 100644 index 00000000000000..ab341f6f5fbfd1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ans_gen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ans_gen_pipeline pipeline T5Transformer from santhoshkolloju +author: John Snow Labs +name: ans_gen_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ans_gen_pipeline` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ans_gen_pipeline_en_5.4.2_3.0_1724055446655.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ans_gen_pipeline_en_5.4.2_3.0_1724055446655.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ans_gen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ans_gen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ans_gen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/santhoshkolloju/ans_gen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_en.md b/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_en.md new file mode 100644 index 00000000000000..5990b441955641 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English arat5_freezed T5Transformer from imen11111 +author: John Snow Labs +name: arat5_freezed +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arat5_freezed` is a English model originally trained by imen11111. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arat5_freezed_en_5.4.2_3.0_1724095575836.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arat5_freezed_en_5.4.2_3.0_1724095575836.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("arat5_freezed","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("arat5_freezed", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arat5_freezed| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/imen11111/araT5-freezed \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_pipeline_en.md new file mode 100644 index 00000000000000..2796327a334367 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-arat5_freezed_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English arat5_freezed_pipeline pipeline T5Transformer from imen11111 +author: John Snow Labs +name: arat5_freezed_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arat5_freezed_pipeline` is a English model originally trained by imen11111. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arat5_freezed_pipeline_en_5.4.2_3.0_1724095696546.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arat5_freezed_pipeline_en_5.4.2_3.0_1724095696546.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("arat5_freezed_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("arat5_freezed_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arat5_freezed_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/imen11111/araT5-freezed + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_en.md b/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_en.md new file mode 100644 index 00000000000000..9d49191c557ea0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English arxivedits_intention_classifier_t5_base_fine_grained T5Transformer from chaojiang06 +author: John Snow Labs +name: arxivedits_intention_classifier_t5_base_fine_grained +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arxivedits_intention_classifier_t5_base_fine_grained` is a English model originally trained by chaojiang06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arxivedits_intention_classifier_t5_base_fine_grained_en_5.4.2_3.0_1724090940063.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arxivedits_intention_classifier_t5_base_fine_grained_en_5.4.2_3.0_1724090940063.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("arxivedits_intention_classifier_t5_base_fine_grained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("arxivedits_intention_classifier_t5_base_fine_grained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arxivedits_intention_classifier_t5_base_fine_grained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|948.5 MB| + +## References + +https://huggingface.co/chaojiang06/arXivEdits-intention-classifier-T5-base-fine-grained \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_pipeline_en.md new file mode 100644 index 00000000000000..eae17b0b7c68f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-arxivedits_intention_classifier_t5_base_fine_grained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English arxivedits_intention_classifier_t5_base_fine_grained_pipeline pipeline T5Transformer from chaojiang06 +author: John Snow Labs +name: arxivedits_intention_classifier_t5_base_fine_grained_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arxivedits_intention_classifier_t5_base_fine_grained_pipeline` is a English model originally trained by chaojiang06. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arxivedits_intention_classifier_t5_base_fine_grained_pipeline_en_5.4.2_3.0_1724091004841.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arxivedits_intention_classifier_t5_base_fine_grained_pipeline_en_5.4.2_3.0_1724091004841.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("arxivedits_intention_classifier_t5_base_fine_grained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("arxivedits_intention_classifier_t5_base_fine_grained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arxivedits_intention_classifier_t5_base_fine_grained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|948.5 MB| + +## References + +https://huggingface.co/chaojiang06/arXivEdits-intention-classifier-T5-base-fine-grained + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_en.md new file mode 100644 index 00000000000000..d70fdfbe037068 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English aspol_coqe_vit5_large_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: aspol_coqe_vit5_large_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`aspol_coqe_vit5_large_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/aspol_coqe_vit5_large_v2_en_5.4.2_3.0_1724099500836.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/aspol_coqe_vit5_large_v2_en_5.4.2_3.0_1724099500836.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("aspol_coqe_vit5_large_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("aspol_coqe_vit5_large_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|aspol_coqe_vit5_large_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/ASPOL_COQE_viT5-large_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_pipeline_en.md new file mode 100644 index 00000000000000..bfb8f1a23440d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-aspol_coqe_vit5_large_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English aspol_coqe_vit5_large_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: aspol_coqe_vit5_large_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`aspol_coqe_vit5_large_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/aspol_coqe_vit5_large_v2_pipeline_en_5.4.2_3.0_1724099686972.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/aspol_coqe_vit5_large_v2_pipeline_en_5.4.2_3.0_1724099686972.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("aspol_coqe_vit5_large_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("aspol_coqe_vit5_large_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|aspol_coqe_vit5_large_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/ASPOL_COQE_viT5-large_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_en.md b/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_en.md new file mode 100644 index 00000000000000..a3d6561448232b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English augmented_nodes_with_edge_label_sentence_level_t5_run2 T5Transformer from sheoran95 +author: John Snow Labs +name: augmented_nodes_with_edge_label_sentence_level_t5_run2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`augmented_nodes_with_edge_label_sentence_level_t5_run2` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/augmented_nodes_with_edge_label_sentence_level_t5_run2_en_5.4.2_3.0_1724048794042.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/augmented_nodes_with_edge_label_sentence_level_t5_run2_en_5.4.2_3.0_1724048794042.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("augmented_nodes_with_edge_label_sentence_level_t5_run2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("augmented_nodes_with_edge_label_sentence_level_t5_run2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|augmented_nodes_with_edge_label_sentence_level_t5_run2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|328.9 MB| + +## References + +https://huggingface.co/sheoran95/augmented_nodes_with_edge_label_sentence_level_T5_run2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline_en.md new file mode 100644 index 00000000000000..6294a13a9ed83c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline_en_5.4.2_3.0_1724048811989.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline_en_5.4.2_3.0_1724048811989.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|augmented_nodes_with_edge_label_sentence_level_t5_run2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|328.9 MB| + +## References + +https://huggingface.co/sheoran95/augmented_nodes_with_edge_label_sentence_level_T5_run2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_en.md b/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_en.md new file mode 100644 index 00000000000000..0624b3d532cfea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English autonlp_parrot_paraphrasing_615317556 T5Transformer from spy24 +author: John Snow Labs +name: autonlp_parrot_paraphrasing_615317556 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autonlp_parrot_paraphrasing_615317556` is a English model originally trained by spy24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autonlp_parrot_paraphrasing_615317556_en_5.4.2_3.0_1724076890104.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autonlp_parrot_paraphrasing_615317556_en_5.4.2_3.0_1724076890104.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("autonlp_parrot_paraphrasing_615317556","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("autonlp_parrot_paraphrasing_615317556", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autonlp_parrot_paraphrasing_615317556| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/spy24/autonlp-parrot_paraphrasing-615317556 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_pipeline_en.md new file mode 100644 index 00000000000000..332988a79a67be --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autonlp_parrot_paraphrasing_615317556_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English autonlp_parrot_paraphrasing_615317556_pipeline pipeline T5Transformer from spy24 +author: John Snow Labs +name: autonlp_parrot_paraphrasing_615317556_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autonlp_parrot_paraphrasing_615317556_pipeline` is a English model originally trained by spy24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autonlp_parrot_paraphrasing_615317556_pipeline_en_5.4.2_3.0_1724076937829.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autonlp_parrot_paraphrasing_615317556_pipeline_en_5.4.2_3.0_1724076937829.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("autonlp_parrot_paraphrasing_615317556_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("autonlp_parrot_paraphrasing_615317556_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autonlp_parrot_paraphrasing_615317556_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/spy24/autonlp-parrot_paraphrasing-615317556 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_en.md b/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_en.md new file mode 100644 index 00000000000000..1bdbfa0d988f1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English autonlp_us_tonga_tonga_islands_aus3_606917136 T5Transformer from spy24 +author: John Snow Labs +name: autonlp_us_tonga_tonga_islands_aus3_606917136 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autonlp_us_tonga_tonga_islands_aus3_606917136` is a English model originally trained by spy24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autonlp_us_tonga_tonga_islands_aus3_606917136_en_5.4.2_3.0_1724036209586.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autonlp_us_tonga_tonga_islands_aus3_606917136_en_5.4.2_3.0_1724036209586.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("autonlp_us_tonga_tonga_islands_aus3_606917136","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("autonlp_us_tonga_tonga_islands_aus3_606917136", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autonlp_us_tonga_tonga_islands_aus3_606917136| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/spy24/autonlp-US-to-AUS3-606917136 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline_en.md new file mode 100644 index 00000000000000..ac16f551cb71aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline pipeline T5Transformer from spy24 +author: John Snow Labs +name: autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline` is a English model originally trained by spy24. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline_en_5.4.2_3.0_1724036257619.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline_en_5.4.2_3.0_1724036257619.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autonlp_us_tonga_tonga_islands_aus3_606917136_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/spy24/autonlp-US-to-AUS3-606917136 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_en.md b/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_en.md new file mode 100644 index 00000000000000..b35d402283b7ec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English autotrain_phrasinator_reverse_670319725 T5Transformer from ianMconversica +author: John Snow Labs +name: autotrain_phrasinator_reverse_670319725 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autotrain_phrasinator_reverse_670319725` is a English model originally trained by ianMconversica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autotrain_phrasinator_reverse_670319725_en_5.4.2_3.0_1724074448659.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autotrain_phrasinator_reverse_670319725_en_5.4.2_3.0_1724074448659.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("autotrain_phrasinator_reverse_670319725","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("autotrain_phrasinator_reverse_670319725", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autotrain_phrasinator_reverse_670319725| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ianMconversica/autotrain-phrasinator-reverse-670319725 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_pipeline_en.md new file mode 100644 index 00000000000000..6a86db4d8d4f63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-autotrain_phrasinator_reverse_670319725_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English autotrain_phrasinator_reverse_670319725_pipeline pipeline T5Transformer from ianMconversica +author: John Snow Labs +name: autotrain_phrasinator_reverse_670319725_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`autotrain_phrasinator_reverse_670319725_pipeline` is a English model originally trained by ianMconversica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/autotrain_phrasinator_reverse_670319725_pipeline_en_5.4.2_3.0_1724074497112.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/autotrain_phrasinator_reverse_670319725_pipeline_en_5.4.2_3.0_1724074497112.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("autotrain_phrasinator_reverse_670319725_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("autotrain_phrasinator_reverse_670319725_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|autotrain_phrasinator_reverse_670319725_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ianMconversica/autotrain-phrasinator-reverse-670319725 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_en.md b/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_en.md new file mode 100644 index 00000000000000..09ef2802cd23a6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bangla_para_v3_390000 T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_390000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_390000` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_390000_en_5.4.2_3.0_1724031946178.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_390000_en_5.4.2_3.0_1724031946178.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bangla_para_v3_390000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bangla_para_v3_390000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_390000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-390000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_pipeline_en.md new file mode 100644 index 00000000000000..e630df6fd1a2a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bangla_para_v3_390000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bangla_para_v3_390000_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_390000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_390000_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_390000_pipeline_en_5.4.2_3.0_1724031992848.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_390000_pipeline_en_5.4.2_3.0_1724031992848.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bangla_para_v3_390000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bangla_para_v3_390000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_390000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-390000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_en.md b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_en.md new file mode 100644 index 00000000000000..3694b25f008df6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bdp_summarization_t10 T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t10` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t10_en_5.4.2_3.0_1724034417705.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t10_en_5.4.2_3.0_1724034417705.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bdp_summarization_t10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bdp_summarization_t10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_pipeline_en.md new file mode 100644 index 00000000000000..c3280f8d46bfda --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bdp_summarization_t10_pipeline pipeline T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t10_pipeline` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t10_pipeline_en_5.4.2_3.0_1724034433328.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t10_pipeline_en_5.4.2_3.0_1724034433328.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bdp_summarization_t10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bdp_summarization_t10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_en.md new file mode 100644 index 00000000000000..a331c48567ee67 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bdp_summarization_t5 T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t5` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t5_en_5.4.2_3.0_1724102301885.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t5_en_5.4.2_3.0_1724102301885.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bdp_summarization_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bdp_summarization_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|338.5 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_pipeline_en.md new file mode 100644 index 00000000000000..7db969720dd725 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bdp_summarization_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bdp_summarization_t5_pipeline pipeline T5Transformer from DamianoDiPaola +author: John Snow Labs +name: bdp_summarization_t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bdp_summarization_t5_pipeline` is a English model originally trained by DamianoDiPaola. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bdp_summarization_t5_pipeline_en_5.4.2_3.0_1724102320662.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bdp_summarization_t5_pipeline_en_5.4.2_3.0_1724102320662.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bdp_summarization_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bdp_summarization_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bdp_summarization_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|338.5 MB| + +## References + +https://huggingface.co/DamianoDiPaola/bdp_summarization_T5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_en.md new file mode 100644 index 00000000000000..f0f68de2a1b9a2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bikes_t5_small T5Transformer from krkv +author: John Snow Labs +name: bikes_t5_small +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small` is a English model originally trained by krkv. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_en_5.4.2_3.0_1724075789441.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_en_5.4.2_3.0_1724075789441.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bikes_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bikes_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.8 MB| + +## References + +https://huggingface.co/krkv/bikes-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..8d0fe76b5ecf32 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-bikes_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bikes_t5_small_pipeline pipeline T5Transformer from krkv +author: John Snow Labs +name: bikes_t5_small_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small_pipeline` is a English model originally trained by krkv. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_pipeline_en_5.4.2_3.0_1724075807640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_pipeline_en_5.4.2_3.0_1724075807640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bikes_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bikes_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.8 MB| + +## References + +https://huggingface.co/krkv/bikes-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_en.md b/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_en.md new file mode 100644 index 00000000000000..cc464c52a012a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English biot5_base_dti_human T5Transformer from QizhiPei +author: John Snow Labs +name: biot5_base_dti_human +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`biot5_base_dti_human` is a English model originally trained by QizhiPei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/biot5_base_dti_human_en_5.4.2_3.0_1724064561870.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/biot5_base_dti_human_en_5.4.2_3.0_1724064561870.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("biot5_base_dti_human","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("biot5_base_dti_human", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|biot5_base_dti_human| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/QizhiPei/biot5-base-dti-human \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_pipeline_en.md new file mode 100644 index 00000000000000..a8da98c83edf8e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-biot5_base_dti_human_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English biot5_base_dti_human_pipeline pipeline T5Transformer from QizhiPei +author: John Snow Labs +name: biot5_base_dti_human_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`biot5_base_dti_human_pipeline` is a English model originally trained by QizhiPei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/biot5_base_dti_human_pipeline_en_5.4.2_3.0_1724064612031.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/biot5_base_dti_human_pipeline_en_5.4.2_3.0_1724064612031.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("biot5_base_dti_human_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("biot5_base_dti_human_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|biot5_base_dti_human_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/QizhiPei/biot5-base-dti-human + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_en.md new file mode 100644 index 00000000000000..7c3909cbfef2fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_jennndexter T5Transformer from JennnDexter +author: John Snow Labs +name: burmese_awesome_billsum_model_jennndexter +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_jennndexter` is a English model originally trained by JennnDexter. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_jennndexter_en_5.4.2_3.0_1724086482512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_jennndexter_en_5.4.2_3.0_1724086482512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_jennndexter","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_jennndexter", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_jennndexter| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|321.7 MB| + +## References + +https://huggingface.co/JennnDexter/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_pipeline_en.md new file mode 100644 index 00000000000000..eb33942032932d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_jennndexter_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_jennndexter_pipeline pipeline T5Transformer from JennnDexter +author: John Snow Labs +name: burmese_awesome_billsum_model_jennndexter_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_jennndexter_pipeline` is a English model originally trained by JennnDexter. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_jennndexter_pipeline_en_5.4.2_3.0_1724086523962.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_jennndexter_pipeline_en_5.4.2_3.0_1724086523962.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_jennndexter_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_jennndexter_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_jennndexter_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|321.7 MB| + +## References + +https://huggingface.co/JennnDexter/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_en.md new file mode 100644 index 00000000000000..b39a4cf35bb5c5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mprimus T5Transformer from Mprimus +author: John Snow Labs +name: burmese_awesome_billsum_model_mprimus +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mprimus` is a English model originally trained by Mprimus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mprimus_en_5.4.2_3.0_1724097453040.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mprimus_en_5.4.2_3.0_1724097453040.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mprimus","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_billsum_model_mprimus", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mprimus| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/Mprimus/my_awesome_billsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_pipeline_en.md new file mode 100644 index 00000000000000..c3e03cf20cc39a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_billsum_model_mprimus_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_billsum_model_mprimus_pipeline pipeline T5Transformer from Mprimus +author: John Snow Labs +name: burmese_awesome_billsum_model_mprimus_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_billsum_model_mprimus_pipeline` is a English model originally trained by Mprimus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mprimus_pipeline_en_5.4.2_3.0_1724097477038.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_billsum_model_mprimus_pipeline_en_5.4.2_3.0_1724097477038.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_billsum_model_mprimus_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_billsum_model_mprimus_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_billsum_model_mprimus_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/Mprimus/my_awesome_billsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_en.md new file mode 100644 index 00000000000000..f884270a6b0d69 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_gingun18 T5Transformer from gingun18 +author: John Snow Labs +name: burmese_awesome_opus_books_model_gingun18 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_gingun18` is a English model originally trained by gingun18. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_gingun18_en_5.4.2_3.0_1724040800455.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_gingun18_en_5.4.2_3.0_1724040800455.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_gingun18","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_gingun18", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_gingun18| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.3 MB| + +## References + +https://huggingface.co/gingun18/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_pipeline_en.md new file mode 100644 index 00000000000000..17927ac4947985 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_gingun18_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_gingun18_pipeline pipeline T5Transformer from gingun18 +author: John Snow Labs +name: burmese_awesome_opus_books_model_gingun18_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_gingun18_pipeline` is a English model originally trained by gingun18. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_gingun18_pipeline_en_5.4.2_3.0_1724040817870.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_gingun18_pipeline_en_5.4.2_3.0_1724040817870.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_gingun18_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_gingun18_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_gingun18_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/gingun18/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_en.md new file mode 100644 index 00000000000000..0899e6f94eb5ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_tippawan T5Transformer from Tippawan +author: John Snow Labs +name: burmese_awesome_opus_books_model_tippawan +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_tippawan` is a English model originally trained by Tippawan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_tippawan_en_5.4.2_3.0_1724057076807.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_tippawan_en_5.4.2_3.0_1724057076807.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_tippawan","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_tippawan", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_tippawan| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.3 MB| + +## References + +https://huggingface.co/Tippawan/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_pipeline_en.md new file mode 100644 index 00000000000000..27dd10c50c20e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_tippawan_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_tippawan_pipeline pipeline T5Transformer from Tippawan +author: John Snow Labs +name: burmese_awesome_opus_books_model_tippawan_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_tippawan_pipeline` is a English model originally trained by Tippawan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_tippawan_pipeline_en_5.4.2_3.0_1724057094534.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_tippawan_pipeline_en_5.4.2_3.0_1724057094534.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_tippawan_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_tippawan_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_tippawan_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.3 MB| + +## References + +https://huggingface.co/Tippawan/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_en.md new file mode 100644 index 00000000000000..7bdbe552b15f22 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_turboquant T5Transformer from turboquant +author: John Snow Labs +name: burmese_awesome_opus_books_model_turboquant +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_turboquant` is a English model originally trained by turboquant. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_turboquant_en_5.4.2_3.0_1724066898788.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_turboquant_en_5.4.2_3.0_1724066898788.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_turboquant","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_turboquant", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_turboquant| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.0 MB| + +## References + +https://huggingface.co/turboquant/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_pipeline_en.md new file mode 100644 index 00000000000000..946e1339d4e40c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_turboquant_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_turboquant_pipeline pipeline T5Transformer from turboquant +author: John Snow Labs +name: burmese_awesome_opus_books_model_turboquant_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_turboquant_pipeline` is a English model originally trained by turboquant. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_turboquant_pipeline_en_5.4.2_3.0_1724066917208.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_turboquant_pipeline_en_5.4.2_3.0_1724066917208.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_turboquant_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_turboquant_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_turboquant_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.0 MB| + +## References + +https://huggingface.co/turboquant/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_en.md new file mode 100644 index 00000000000000..ac41340f9a0afc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_williamstar T5Transformer from WilliamStar +author: John Snow Labs +name: burmese_awesome_opus_books_model_williamstar +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_williamstar` is a English model originally trained by WilliamStar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_williamstar_en_5.4.2_3.0_1724096836271.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_williamstar_en_5.4.2_3.0_1724096836271.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_williamstar","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_williamstar", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_williamstar| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.8 MB| + +## References + +https://huggingface.co/WilliamStar/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_pipeline_en.md new file mode 100644 index 00000000000000..933666b206bbad --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-burmese_awesome_opus_books_model_williamstar_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_williamstar_pipeline pipeline T5Transformer from WilliamStar +author: John Snow Labs +name: burmese_awesome_opus_books_model_williamstar_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_williamstar_pipeline` is a English model originally trained by WilliamStar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_williamstar_pipeline_en_5.4.2_3.0_1724096854038.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_williamstar_pipeline_en_5.4.2_3.0_1724096854038.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_williamstar_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_williamstar_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_williamstar_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.8 MB| + +## References + +https://huggingface.co/WilliamStar/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_en.md b/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_en.md new file mode 100644 index 00000000000000..392342b43eb2f5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cabecalho_german_ementas_com_ptt5 T5Transformer from llmf +author: John Snow Labs +name: cabecalho_german_ementas_com_ptt5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cabecalho_german_ementas_com_ptt5` is a English model originally trained by llmf. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cabecalho_german_ementas_com_ptt5_en_5.4.2_3.0_1724061338308.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cabecalho_german_ementas_com_ptt5_en_5.4.2_3.0_1724061338308.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cabecalho_german_ementas_com_ptt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cabecalho_german_ementas_com_ptt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cabecalho_german_ementas_com_ptt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|953.3 MB| + +## References + +https://huggingface.co/llmf/cabecalho-de-ementas-com-ptt5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_pipeline_en.md new file mode 100644 index 00000000000000..3855acb225bdae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cabecalho_german_ementas_com_ptt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cabecalho_german_ementas_com_ptt5_pipeline pipeline T5Transformer from llmf +author: John Snow Labs +name: cabecalho_german_ementas_com_ptt5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cabecalho_german_ementas_com_ptt5_pipeline` is a English model originally trained by llmf. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cabecalho_german_ementas_com_ptt5_pipeline_en_5.4.2_3.0_1724061397234.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cabecalho_german_ementas_com_ptt5_pipeline_en_5.4.2_3.0_1724061397234.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cabecalho_german_ementas_com_ptt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cabecalho_german_ementas_com_ptt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cabecalho_german_ementas_com_ptt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|953.3 MB| + +## References + +https://huggingface.co/llmf/cabecalho-de-ementas-com-ptt5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_en.md b/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_en.md new file mode 100644 index 00000000000000..a2e9159c31b2b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English carrier_progression T5Transformer from ManulaPankaja +author: John Snow Labs +name: carrier_progression +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`carrier_progression` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/carrier_progression_en_5.4.2_3.0_1724096067216.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/carrier_progression_en_5.4.2_3.0_1724096067216.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("carrier_progression","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("carrier_progression", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|carrier_progression| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|973.0 MB| + +## References + +https://huggingface.co/ManulaPankaja/carrier_progression \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_pipeline_en.md new file mode 100644 index 00000000000000..d93ad7ae56f5ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-carrier_progression_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English carrier_progression_pipeline pipeline T5Transformer from ManulaPankaja +author: John Snow Labs +name: carrier_progression_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`carrier_progression_pipeline` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/carrier_progression_pipeline_en_5.4.2_3.0_1724096123749.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/carrier_progression_pipeline_en_5.4.2_3.0_1724096123749.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("carrier_progression_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("carrier_progression_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|carrier_progression_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|973.0 MB| + +## References + +https://huggingface.co/ManulaPankaja/carrier_progression + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-classification_test_en.md b/docs/_posts/ahmedlone127/2024-08-19-classification_test_en.md new file mode 100644 index 00000000000000..f14df31191c305 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-classification_test_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English classification_test T5Transformer from ahsan-mavros +author: John Snow Labs +name: classification_test +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`classification_test` is a English model originally trained by ahsan-mavros. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/classification_test_en_5.4.2_3.0_1724066062868.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/classification_test_en_5.4.2_3.0_1724066062868.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("classification_test","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("classification_test", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|classification_test| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ahsan-mavros/classification-test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-classification_test_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-classification_test_pipeline_en.md new file mode 100644 index 00000000000000..c3140aa5181486 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-classification_test_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English classification_test_pipeline pipeline T5Transformer from ahsan-mavros +author: John Snow Labs +name: classification_test_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`classification_test_pipeline` is a English model originally trained by ahsan-mavros. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/classification_test_pipeline_en_5.4.2_3.0_1724066111990.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/classification_test_pipeline_en_5.4.2_3.0_1724066111990.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("classification_test_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("classification_test_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|classification_test_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ahsan-mavros/classification-test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_en.md b/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_en.md new file mode 100644 index 00000000000000..7b0dfd605d1af7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cn_maltese_small T5Transformer from arfu +author: John Snow Labs +name: cn_maltese_small +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cn_maltese_small` is a English model originally trained by arfu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cn_maltese_small_en_5.4.2_3.0_1724053563412.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cn_maltese_small_en_5.4.2_3.0_1724053563412.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cn_maltese_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cn_maltese_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cn_maltese_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|353.9 MB| + +## References + +https://huggingface.co/arfu/cn-mt-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_pipeline_en.md new file mode 100644 index 00000000000000..ee768e555a8ef9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cn_maltese_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cn_maltese_small_pipeline pipeline T5Transformer from arfu +author: John Snow Labs +name: cn_maltese_small_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cn_maltese_small_pipeline` is a English model originally trained by arfu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cn_maltese_small_pipeline_en_5.4.2_3.0_1724053583081.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cn_maltese_small_pipeline_en_5.4.2_3.0_1724053583081.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cn_maltese_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cn_maltese_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cn_maltese_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|353.9 MB| + +## References + +https://huggingface.co/arfu/cn-mt-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_en.md new file mode 100644 index 00000000000000..c83a4c27bcf102 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English code_corrector_model T5Transformer from manahil1 +author: John Snow Labs +name: code_corrector_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`code_corrector_model` is a English model originally trained by manahil1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/code_corrector_model_en_5.4.2_3.0_1724036203445.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/code_corrector_model_en_5.4.2_3.0_1724036203445.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("code_corrector_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("code_corrector_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|code_corrector_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|319.9 MB| + +## References + +https://huggingface.co/manahil1/Code_Corrector_Model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_pipeline_en.md new file mode 100644 index 00000000000000..5c0c6fd869b4df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-code_corrector_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English code_corrector_model_pipeline pipeline T5Transformer from manahil1 +author: John Snow Labs +name: code_corrector_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`code_corrector_model_pipeline` is a English model originally trained by manahil1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/code_corrector_model_pipeline_en_5.4.2_3.0_1724036224313.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/code_corrector_model_pipeline_en_5.4.2_3.0_1724036224313.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("code_corrector_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("code_corrector_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|code_corrector_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|319.9 MB| + +## References + +https://huggingface.co/manahil1/Code_Corrector_Model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_en.md b/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_en.md new file mode 100644 index 00000000000000..55c6143dfc4a76 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English code_mixed_banglish_english_3 T5Transformer from Ayon128 +author: John Snow Labs +name: code_mixed_banglish_english_3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`code_mixed_banglish_english_3` is a English model originally trained by Ayon128. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/code_mixed_banglish_english_3_en_5.4.2_3.0_1724090683446.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/code_mixed_banglish_english_3_en_5.4.2_3.0_1724090683446.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("code_mixed_banglish_english_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("code_mixed_banglish_english_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|code_mixed_banglish_english_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Ayon128/code-mixed_Banglish_English_3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_pipeline_en.md new file mode 100644 index 00000000000000..993539206558fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-code_mixed_banglish_english_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English code_mixed_banglish_english_3_pipeline pipeline T5Transformer from Ayon128 +author: John Snow Labs +name: code_mixed_banglish_english_3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`code_mixed_banglish_english_3_pipeline` is a English model originally trained by Ayon128. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/code_mixed_banglish_english_3_pipeline_en_5.4.2_3.0_1724090729974.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/code_mixed_banglish_english_3_pipeline_en_5.4.2_3.0_1724090729974.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("code_mixed_banglish_english_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("code_mixed_banglish_english_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|code_mixed_banglish_english_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Ayon128/code-mixed_Banglish_English_3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_en.md new file mode 100644 index 00000000000000..8bba98ee379d14 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English commongen_t5_base T5Transformer from guyhadad01 +author: John Snow Labs +name: commongen_t5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`commongen_t5_base` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/commongen_t5_base_en_5.4.2_3.0_1724063246496.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/commongen_t5_base_en_5.4.2_3.0_1724063246496.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("commongen_t5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("commongen_t5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|commongen_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/guyhadad01/commongen-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_pipeline_en.md new file mode 100644 index 00000000000000..550d74d071303b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-commongen_t5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English commongen_t5_base_pipeline pipeline T5Transformer from guyhadad01 +author: John Snow Labs +name: commongen_t5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`commongen_t5_base_pipeline` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/commongen_t5_base_pipeline_en_5.4.2_3.0_1724063299588.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/commongen_t5_base_pipeline_en_5.4.2_3.0_1724063299588.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("commongen_t5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("commongen_t5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|commongen_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/guyhadad01/commongen-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_en.md b/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_en.md new file mode 100644 index 00000000000000..686ab32972fff8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English corrupt_grammar_model_wmt460 T5Transformer from mika5883 +author: John Snow Labs +name: corrupt_grammar_model_wmt460 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`corrupt_grammar_model_wmt460` is a English model originally trained by mika5883. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/corrupt_grammar_model_wmt460_en_5.4.2_3.0_1724060456404.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/corrupt_grammar_model_wmt460_en_5.4.2_3.0_1724060456404.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("corrupt_grammar_model_wmt460","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("corrupt_grammar_model_wmt460", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|corrupt_grammar_model_wmt460| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mika5883/corrupt_grammar_model_wmt460 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_pipeline_en.md new file mode 100644 index 00000000000000..31fe0c12dcd613 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-corrupt_grammar_model_wmt460_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English corrupt_grammar_model_wmt460_pipeline pipeline T5Transformer from mika5883 +author: John Snow Labs +name: corrupt_grammar_model_wmt460_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`corrupt_grammar_model_wmt460_pipeline` is a English model originally trained by mika5883. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/corrupt_grammar_model_wmt460_pipeline_en_5.4.2_3.0_1724060505645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/corrupt_grammar_model_wmt460_pipeline_en_5.4.2_3.0_1724060505645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("corrupt_grammar_model_wmt460_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("corrupt_grammar_model_wmt460_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|corrupt_grammar_model_wmt460_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mika5883/corrupt_grammar_model_wmt460 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_en.md b/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_en.md new file mode 100644 index 00000000000000..7018bb2a1f6251 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cpt_kginfusedlm_sakharam_aviation_triples20 T5Transformer from sakharamg +author: John Snow Labs +name: cpt_kginfusedlm_sakharam_aviation_triples20 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cpt_kginfusedlm_sakharam_aviation_triples20` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cpt_kginfusedlm_sakharam_aviation_triples20_en_5.4.2_3.0_1724040391603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cpt_kginfusedlm_sakharam_aviation_triples20_en_5.4.2_3.0_1724040391603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cpt_kginfusedlm_sakharam_aviation_triples20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cpt_kginfusedlm_sakharam_aviation_triples20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cpt_kginfusedlm_sakharam_aviation_triples20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/CPT_KGinfusedLM_sakharam_aviation_triples20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_pipeline_en.md new file mode 100644 index 00000000000000..5512a837de3011 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cpt_kginfusedlm_sakharam_aviation_triples20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cpt_kginfusedlm_sakharam_aviation_triples20_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: cpt_kginfusedlm_sakharam_aviation_triples20_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cpt_kginfusedlm_sakharam_aviation_triples20_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cpt_kginfusedlm_sakharam_aviation_triples20_pipeline_en_5.4.2_3.0_1724040525348.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cpt_kginfusedlm_sakharam_aviation_triples20_pipeline_en_5.4.2_3.0_1724040525348.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cpt_kginfusedlm_sakharam_aviation_triples20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cpt_kginfusedlm_sakharam_aviation_triples20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cpt_kginfusedlm_sakharam_aviation_triples20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/CPT_KGinfusedLM_sakharam_aviation_triples20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_en.md new file mode 100644 index 00000000000000..c2f91e83458c82 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_ospal T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_ospal +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_ospal` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_ospal_en_5.4.2_3.0_1724064765900.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_ospal_en_5.4.2_3.0_1724064765900.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_ospal","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_ospal", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_ospal| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_OSPAL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_pipeline_en.md new file mode 100644 index 00000000000000..1f8a192ab7d9ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_prompting5_ospal_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_ospal_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_ospal_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_ospal_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_ospal_pipeline_en_5.4.2_3.0_1724064939830.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_ospal_pipeline_en_5.4.2_3.0_1724064939830.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_ospal_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_ospal_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_ospal_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_OSPAL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_aopsl_v1_h0_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_aopsl_v1_h0_en.md new file mode 100644 index 00000000000000..ad6e2e3c0f2413 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_aopsl_v1_h0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_aopsl_v1_h0 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_aopsl_v1_h0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_aopsl_v1_h0` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aopsl_v1_h0_en_5.4.2_3.0_1724107401885.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aopsl_v1_h0_en_5.4.2_3.0_1724107401885.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_aopsl_v1_h0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_aopsl_v1_h0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_aopsl_v1_h0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_AOPSL_v1_h0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_en.md new file mode 100644 index 00000000000000..519c07ea6dd84b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_saopl_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_saopl_v1_h1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_saopl_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_h1_en_5.4.2_3.0_1724110408093.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_h1_en_5.4.2_3.0_1724110408093.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_saopl_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_saopl_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_saopl_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAOPL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..9803d4308dcdc6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline_en_5.4.2_3.0_1724110579671.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline_en_5.4.2_3.0_1724110579671.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_saopl_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAOPL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_en.md new file mode 100644 index 00000000000000..47e5b57f3f993b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_sapol_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_sapol_v1_h1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_sapol_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sapol_v1_h1_en_5.4.2_3.0_1724067733919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sapol_v1_h1_en_5.4.2_3.0_1724067733919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_sapol_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_sapol_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_sapol_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SAPOL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..243fc2dc0f2d73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline_en_5.4.2_3.0_1724067916139.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline_en_5.4.2_3.0_1724067916139.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_sapol_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_SAPOL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_en.md new file mode 100644 index 00000000000000..523bbd361b18a8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction3_soapl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction3_soapl_v1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction3_soapl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_v1_en_5.4.2_3.0_1724082738938.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_v1_en_5.4.2_3.0_1724082738938.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction3_soapl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction3_soapl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction3_soapl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction3_SOAPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline_en.md new file mode 100644 index 00000000000000..ad585ad5382f98 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline_en_5.4.2_3.0_1724082933235.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline_en_5.4.2_3.0_1724082933235.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction3_soapl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction3_SOAPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_en.md new file mode 100644 index 00000000000000..07aecb46a4f354 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cussupport_chat_t5_small T5Transformer from mrSoul7766 +author: John Snow Labs +name: cussupport_chat_t5_small +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cussupport_chat_t5_small` is a English model originally trained by mrSoul7766. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cussupport_chat_t5_small_en_5.4.2_3.0_1724049588195.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cussupport_chat_t5_small_en_5.4.2_3.0_1724049588195.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cussupport_chat_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cussupport_chat_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cussupport_chat_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|346.8 MB| + +## References + +https://huggingface.co/mrSoul7766/CUSsupport-chat-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..c2a23f56e5d622 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-cussupport_chat_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cussupport_chat_t5_small_pipeline pipeline T5Transformer from mrSoul7766 +author: John Snow Labs +name: cussupport_chat_t5_small_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cussupport_chat_t5_small_pipeline` is a English model originally trained by mrSoul7766. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cussupport_chat_t5_small_pipeline_en_5.4.2_3.0_1724049605386.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cussupport_chat_t5_small_pipeline_en_5.4.2_3.0_1724049605386.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cussupport_chat_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cussupport_chat_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cussupport_chat_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|346.8 MB| + +## References + +https://huggingface.co/mrSoul7766/CUSsupport-chat-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_en.md b/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_en.md new file mode 100644 index 00000000000000..f1f4fd0d5948a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dfm_cosql T5Transformer from donmaclean +author: John Snow Labs +name: dfm_cosql +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dfm_cosql` is a English model originally trained by donmaclean. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dfm_cosql_en_5.4.2_3.0_1724099628468.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dfm_cosql_en_5.4.2_3.0_1724099628468.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dfm_cosql","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dfm_cosql", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dfm_cosql| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/donmaclean/dfm_cosql \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_pipeline_en.md new file mode 100644 index 00000000000000..d12530a8de23f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-dfm_cosql_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English dfm_cosql_pipeline pipeline T5Transformer from donmaclean +author: John Snow Labs +name: dfm_cosql_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dfm_cosql_pipeline` is a English model originally trained by donmaclean. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dfm_cosql_pipeline_en_5.4.2_3.0_1724099683773.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dfm_cosql_pipeline_en_5.4.2_3.0_1724099683773.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("dfm_cosql_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("dfm_cosql_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dfm_cosql_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/donmaclean/dfm_cosql + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_en.md new file mode 100644 index 00000000000000..fa6dab64275672 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_0_07_0_25 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_07_0_25 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_07_0_25` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_07_0_25_en_5.4.2_3.0_1724101783803.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_07_0_25_en_5.4.2_3.0_1724101783803.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_0_07_0_25","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_0_07_0_25", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_07_0_25| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.07-0.25 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_pipeline_en.md new file mode 100644 index 00000000000000..68e505eafbf841 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_0_07_0_25_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_0_07_0_25_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_07_0_25_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_07_0_25_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_07_0_25_pipeline_en_5.4.2_3.0_1724101974606.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_07_0_25_pipeline_en_5.4.2_3.0_1724101974606.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_0_07_0_25_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_0_07_0_25_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_07_0_25_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.07-0.25 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_en.md new file mode 100644 index 00000000000000..d59e100684a245 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_1_1 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_1_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_1_1` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_1_1_en_5.4.2_3.0_1724033946203.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_1_1_en_5.4.2_3.0_1724033946203.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_1_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_1_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_1_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-1-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_pipeline_en.md new file mode 100644 index 00000000000000..9f393592426aab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_1_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_1_1_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_1_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_1_1_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_1_1_pipeline_en_5.4.2_3.0_1724034124621.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_1_1_pipeline_en_5.4.2_3.0_1724034124621.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_1_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_1_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_1_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-1-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_en.md new file mode 100644 index 00000000000000..c735f2b3562bb6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_b0_02 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b0_02 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b0_02` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b0_02_en_5.4.2_3.0_1724084364301.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b0_02_en_5.4.2_3.0_1724084364301.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_b0_02","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_b0_02", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b0_02| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b0.02 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_pipeline_en.md new file mode 100644 index 00000000000000..7c7af179dac84f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b0_02_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_b0_02_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b0_02_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b0_02_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b0_02_pipeline_en_5.4.2_3.0_1724084552470.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b0_02_pipeline_en_5.4.2_3.0_1724084552470.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_b0_02_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_b0_02_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b0_02_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b0.02 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_en.md new file mode 100644 index 00000000000000..49f0bd45321c5d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_b2 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b2` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b2_en_5.4.2_3.0_1724084829541.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b2_en_5.4.2_3.0_1724084829541.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_b2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_b2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_pipeline_en.md new file mode 100644 index 00000000000000..414368aff7e882 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_b2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_b2_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b2_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b2_pipeline_en_5.4.2_3.0_1724085020554.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b2_pipeline_en_5.4.2_3.0_1724085020554.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_b2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_b2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_en.md new file mode 100644 index 00000000000000..a0a392cf1e71ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_test T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_test +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_test` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_test_en_5.4.2_3.0_1724066405183.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_test_en_5.4.2_3.0_1724066405183.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_test","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_test", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_test| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_pipeline_en.md new file mode 100644 index 00000000000000..4a3814630173bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilled_mt5_small_test_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_test_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_test_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_test_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_test_pipeline_en_5.4.2_3.0_1724066596770.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_test_pipeline_en_5.4.2_3.0_1724066596770.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_test_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_test_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_test_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_en.md new file mode 100644 index 00000000000000..2f7a9034747182 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilt5_qg_hl_6_4 T5Transformer from valhalla +author: John Snow Labs +name: distilt5_qg_hl_6_4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilt5_qg_hl_6_4` is a English model originally trained by valhalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilt5_qg_hl_6_4_en_5.4.2_3.0_1724063979709.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilt5_qg_hl_6_4_en_5.4.2_3.0_1724063979709.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilt5_qg_hl_6_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilt5_qg_hl_6_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilt5_qg_hl_6_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|317.0 MB| + +## References + +https://huggingface.co/valhalla/distilt5-qg-hl-6-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_pipeline_en.md new file mode 100644 index 00000000000000..1eefac7e4ec807 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-distilt5_qg_hl_6_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilt5_qg_hl_6_4_pipeline pipeline T5Transformer from valhalla +author: John Snow Labs +name: distilt5_qg_hl_6_4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilt5_qg_hl_6_4_pipeline` is a English model originally trained by valhalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilt5_qg_hl_6_4_pipeline_en_5.4.2_3.0_1724063994841.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilt5_qg_hl_6_4_pipeline_en_5.4.2_3.0_1724063994841.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilt5_qg_hl_6_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilt5_qg_hl_6_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilt5_qg_hl_6_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|317.0 MB| + +## References + +https://huggingface.co/valhalla/distilt5-qg-hl-6-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-dream_t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-dream_t5_en.md new file mode 100644 index 00000000000000..2520cdf7fd3bd3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-dream_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dream_t5 T5Transformer from gustavecortal +author: John Snow Labs +name: dream_t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dream_t5` is a English model originally trained by gustavecortal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dream_t5_en_5.4.2_3.0_1724050064527.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dream_t5_en_5.4.2_3.0_1724050064527.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dream_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dream_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dream_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/gustavecortal/dream-t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-dream_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-dream_t5_pipeline_en.md new file mode 100644 index 00000000000000..fb087ddecc1a24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-dream_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English dream_t5_pipeline pipeline T5Transformer from gustavecortal +author: John Snow Labs +name: dream_t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dream_t5_pipeline` is a English model originally trained by gustavecortal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dream_t5_pipeline_en_5.4.2_3.0_1724050111354.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dream_t5_pipeline_en_5.4.2_3.0_1724050111354.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("dream_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("dream_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dream_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/gustavecortal/dream-t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-drone_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-drone_2_en.md new file mode 100644 index 00000000000000..dbeb70e50f8d01 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-drone_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English drone_2 T5Transformer from sahithya20 +author: John Snow Labs +name: drone_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`drone_2` is a English model originally trained by sahithya20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/drone_2_en_5.4.2_3.0_1724036279076.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/drone_2_en_5.4.2_3.0_1724036279076.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("drone_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("drone_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|drone_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|305.9 MB| + +## References + +https://huggingface.co/sahithya20/drone_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-drone_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-drone_2_pipeline_en.md new file mode 100644 index 00000000000000..909ef2b6ca4aba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-drone_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English drone_2_pipeline pipeline T5Transformer from sahithya20 +author: John Snow Labs +name: drone_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`drone_2_pipeline` is a English model originally trained by sahithya20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/drone_2_pipeline_en_5.4.2_3.0_1724036303289.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/drone_2_pipeline_en_5.4.2_3.0_1724036303289.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("drone_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("drone_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|drone_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|305.9 MB| + +## References + +https://huggingface.co/sahithya20/drone_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_en.md b/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_en.md new file mode 100644 index 00000000000000..ebb8a5b771cb1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English end2endqgt5 T5Transformer from BOULLOUL +author: John Snow Labs +name: end2endqgt5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`end2endqgt5` is a English model originally trained by BOULLOUL. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/end2endqgt5_en_5.4.2_3.0_1724097605520.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/end2endqgt5_en_5.4.2_3.0_1724097605520.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("end2endqgt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("end2endqgt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|end2endqgt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/BOULLOUL/End2EndQGT5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_pipeline_en.md new file mode 100644 index 00000000000000..7c5e1f478eb600 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-end2endqgt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English end2endqgt5_pipeline pipeline T5Transformer from BOULLOUL +author: John Snow Labs +name: end2endqgt5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`end2endqgt5_pipeline` is a English model originally trained by BOULLOUL. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/end2endqgt5_pipeline_en_5.4.2_3.0_1724097658442.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/end2endqgt5_pipeline_en_5.4.2_3.0_1724097658442.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("end2endqgt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("end2endqgt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|end2endqgt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/BOULLOUL/End2EndQGT5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_en.md new file mode 100644 index 00000000000000..5568ef888c5a51 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_2_spanish_model_2 T5Transformer from TigerUppercut77 +author: John Snow Labs +name: english_2_spanish_model_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_2_spanish_model_2` is a English model originally trained by TigerUppercut77. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_2_spanish_model_2_en_5.4.2_3.0_1724111981663.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_2_spanish_model_2_en_5.4.2_3.0_1724111981663.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_2_spanish_model_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_2_spanish_model_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_2_spanish_model_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.5 MB| + +## References + +https://huggingface.co/TigerUppercut77/english_2_spanish_model_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_pipeline_en.md new file mode 100644 index 00000000000000..00579ad199136c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_2_spanish_model_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_2_spanish_model_2_pipeline pipeline T5Transformer from TigerUppercut77 +author: John Snow Labs +name: english_2_spanish_model_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_2_spanish_model_2_pipeline` is a English model originally trained by TigerUppercut77. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_2_spanish_model_2_pipeline_en_5.4.2_3.0_1724111999817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_2_spanish_model_2_pipeline_en_5.4.2_3.0_1724111999817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_2_spanish_model_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_2_spanish_model_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_2_spanish_model_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.5 MB| + +## References + +https://huggingface.co/TigerUppercut77/english_2_spanish_model_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_en.md new file mode 100644 index 00000000000000..a10107e653592a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_mt5_base_spider T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_base_spider +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_base_spider` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_base_spider_en_5.4.2_3.0_1724085267674.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_base_spider_en_5.4.2_3.0_1724085267674.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_mt5_base_spider","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_mt5_base_spider", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_base_spider| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-base_spider \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_pipeline_en.md new file mode 100644 index 00000000000000..e288e872b167d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_base_spider_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_mt5_base_spider_pipeline pipeline T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_base_spider_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_base_spider_pipeline` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_base_spider_pipeline_en_5.4.2_3.0_1724085590854.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_base_spider_pipeline_en_5.4.2_3.0_1724085590854.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_mt5_base_spider_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_mt5_base_spider_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_base_spider_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-base_spider + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_en.md new file mode 100644 index 00000000000000..11cbc20c884540 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_mt5_small_10_spider T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_small_10_spider +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_small_10_spider` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_small_10_spider_en_5.4.2_3.0_1724077653478.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_small_10_spider_en_5.4.2_3.0_1724077653478.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_mt5_small_10_spider","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_mt5_small_10_spider", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_small_10_spider| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-small_10_spider \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_pipeline_en.md new file mode 100644 index 00000000000000..0c54d75e385a35 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_mt5_small_10_spider_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_mt5_small_10_spider_pipeline pipeline T5Transformer from e22vvb +author: John Snow Labs +name: english_mt5_small_10_spider_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_mt5_small_10_spider_pipeline` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_mt5_small_10_spider_pipeline_en_5.4.2_3.0_1724077841193.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_mt5_small_10_spider_pipeline_en_5.4.2_3.0_1724077841193.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_mt5_small_10_spider_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_mt5_small_10_spider_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_mt5_small_10_spider_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/e22vvb/EN_mt5-small_10_spider + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-english_vietnamese_mt5_base_docs_news_train_en.md b/docs/_posts/ahmedlone127/2024-08-19-english_vietnamese_mt5_base_docs_news_train_en.md new file mode 100644 index 00000000000000..fd0fe99c39a66c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-english_vietnamese_mt5_base_docs_news_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_vietnamese_mt5_base_docs_news_train T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_mt5_base_docs_news_train +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_mt5_base_docs_news_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_docs_news_train_en_5.4.2_3.0_1724092288707.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_docs_news_train_en_5.4.2_3.0_1724092288707.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_vietnamese_mt5_base_docs_news_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_vietnamese_mt5_base_docs_news_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_mt5_base_docs_news_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_mt5-base_docs_news_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_en.md new file mode 100644 index 00000000000000..73af75e834a35e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_keyword_t5_small_standard_bahasa_cased T5Transformer from mesolitica +author: John Snow Labs +name: finetune_keyword_t5_small_standard_bahasa_cased +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_keyword_t5_small_standard_bahasa_cased` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_small_standard_bahasa_cased_en_5.4.2_3.0_1724045987692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_small_standard_bahasa_cased_en_5.4.2_3.0_1724045987692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_keyword_t5_small_standard_bahasa_cased","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_keyword_t5_small_standard_bahasa_cased", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_keyword_t5_small_standard_bahasa_cased| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.3 MB| + +## References + +https://huggingface.co/mesolitica/finetune-keyword-t5-small-standard-bahasa-cased \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_pipeline_en.md new file mode 100644 index 00000000000000..956a38e5daec13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_keyword_t5_small_standard_bahasa_cased_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_keyword_t5_small_standard_bahasa_cased_pipeline pipeline T5Transformer from mesolitica +author: John Snow Labs +name: finetune_keyword_t5_small_standard_bahasa_cased_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_keyword_t5_small_standard_bahasa_cased_pipeline` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_small_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724046011415.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_small_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724046011415.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_keyword_t5_small_standard_bahasa_cased_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_keyword_t5_small_standard_bahasa_cased_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_keyword_t5_small_standard_bahasa_cased_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.3 MB| + +## References + +https://huggingface.co/mesolitica/finetune-keyword-t5-small-standard-bahasa-cased + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_en.md new file mode 100644 index 00000000000000..cf80a89f46f551 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_t5_small_gazeta_full T5Transformer from rhythm00 +author: John Snow Labs +name: finetune_t5_small_gazeta_full +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_t5_small_gazeta_full` is a English model originally trained by rhythm00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_full_en_5.4.2_3.0_1724097258764.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_full_en_5.4.2_3.0_1724097258764.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_t5_small_gazeta_full","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_t5_small_gazeta_full", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_t5_small_gazeta_full| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|277.9 MB| + +## References + +https://huggingface.co/rhythm00/finetune_t5_small_gazeta_full \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_pipeline_en.md new file mode 100644 index 00000000000000..0e282079a3d0bc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_full_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_t5_small_gazeta_full_pipeline pipeline T5Transformer from rhythm00 +author: John Snow Labs +name: finetune_t5_small_gazeta_full_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_t5_small_gazeta_full_pipeline` is a English model originally trained by rhythm00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_full_pipeline_en_5.4.2_3.0_1724097273620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_full_pipeline_en_5.4.2_3.0_1724097273620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_t5_small_gazeta_full_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_t5_small_gazeta_full_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_t5_small_gazeta_full_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|277.9 MB| + +## References + +https://huggingface.co/rhythm00/finetune_t5_small_gazeta_full + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_en.md new file mode 100644 index 00000000000000..9fdee4b318aed5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_t5_small_gazeta_part T5Transformer from rhythm00 +author: John Snow Labs +name: finetune_t5_small_gazeta_part +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_t5_small_gazeta_part` is a English model originally trained by rhythm00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_part_en_5.4.2_3.0_1724111173180.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_part_en_5.4.2_3.0_1724111173180.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_t5_small_gazeta_part","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_t5_small_gazeta_part", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_t5_small_gazeta_part| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|277.8 MB| + +## References + +https://huggingface.co/rhythm00/finetune_t5_small_gazeta_part \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_pipeline_en.md new file mode 100644 index 00000000000000..70d8bee9242eae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_t5_small_gazeta_part_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_t5_small_gazeta_part_pipeline pipeline T5Transformer from rhythm00 +author: John Snow Labs +name: finetune_t5_small_gazeta_part_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_t5_small_gazeta_part_pipeline` is a English model originally trained by rhythm00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_part_pipeline_en_5.4.2_3.0_1724111188297.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_t5_small_gazeta_part_pipeline_en_5.4.2_3.0_1724111188297.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_t5_small_gazeta_part_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_t5_small_gazeta_part_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_t5_small_gazeta_part_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|277.8 MB| + +## References + +https://huggingface.co/rhythm00/finetune_t5_small_gazeta_part + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_en.md new file mode 100644 index 00000000000000..d8a135ba66e16e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_ttkg_t5_base_standard_bahasa_cased T5Transformer from mesolitica +author: John Snow Labs +name: finetune_ttkg_t5_base_standard_bahasa_cased +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_ttkg_t5_base_standard_bahasa_cased` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_ttkg_t5_base_standard_bahasa_cased_en_5.4.2_3.0_1724080090801.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_ttkg_t5_base_standard_bahasa_cased_en_5.4.2_3.0_1724080090801.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_ttkg_t5_base_standard_bahasa_cased","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_ttkg_t5_base_standard_bahasa_cased", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_ttkg_t5_base_standard_bahasa_cased| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mesolitica/finetune-ttkg-t5-base-standard-bahasa-cased \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_pipeline_en.md new file mode 100644 index 00000000000000..8f93ba3c43baff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_ttkg_t5_base_standard_bahasa_cased_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_ttkg_t5_base_standard_bahasa_cased_pipeline pipeline T5Transformer from mesolitica +author: John Snow Labs +name: finetune_ttkg_t5_base_standard_bahasa_cased_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_ttkg_t5_base_standard_bahasa_cased_pipeline` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_ttkg_t5_base_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724080137385.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_ttkg_t5_base_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724080137385.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_ttkg_t5_base_standard_bahasa_cased_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_ttkg_t5_base_standard_bahasa_cased_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_ttkg_t5_base_standard_bahasa_cased_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mesolitica/finetune-ttkg-t5-base-standard-bahasa-cased + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_en.md new file mode 100644 index 00000000000000..5fdc72e219405e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_zeroshot_ner_t5_small_standard_bahasa_cased T5Transformer from mesolitica +author: John Snow Labs +name: finetune_zeroshot_ner_t5_small_standard_bahasa_cased +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_zeroshot_ner_t5_small_standard_bahasa_cased` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_zeroshot_ner_t5_small_standard_bahasa_cased_en_5.4.2_3.0_1724069422351.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_zeroshot_ner_t5_small_standard_bahasa_cased_en_5.4.2_3.0_1724069422351.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_zeroshot_ner_t5_small_standard_bahasa_cased","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_zeroshot_ner_t5_small_standard_bahasa_cased", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_zeroshot_ner_t5_small_standard_bahasa_cased| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/mesolitica/finetune-zeroshot-ner-t5-small-standard-bahasa-cased \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline_en.md new file mode 100644 index 00000000000000..b34e8d65417165 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline pipeline T5Transformer from mesolitica +author: John Snow Labs +name: finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724069438861.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724069438861.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_zeroshot_ner_t5_small_standard_bahasa_cased_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/mesolitica/finetune-zeroshot-ner-t5-small-standard-bahasa-cased + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_en.md new file mode 100644 index 00000000000000..37b271c90519e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetuned_arat5 T5Transformer from Grinta-king +author: John Snow Labs +name: finetuned_arat5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_arat5` is a English model originally trained by Grinta-king. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_arat5_en_5.4.2_3.0_1724050440319.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_arat5_en_5.4.2_3.0_1724050440319.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetuned_arat5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetuned_arat5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_arat5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/Grinta-king/finetuned-arat5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_pipeline_en.md new file mode 100644 index 00000000000000..4ecb8cfc779675 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetuned_arat5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetuned_arat5_pipeline pipeline T5Transformer from Grinta-king +author: John Snow Labs +name: finetuned_arat5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_arat5_pipeline` is a English model originally trained by Grinta-king. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_arat5_pipeline_en_5.4.2_3.0_1724050516941.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_arat5_pipeline_en_5.4.2_3.0_1724050516941.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetuned_arat5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetuned_arat5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_arat5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/Grinta-king/finetuned-arat5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_en.md new file mode 100644 index 00000000000000..67b0a59bc2d5af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetuned_t5_math T5Transformer from DVLe +author: John Snow Labs +name: finetuned_t5_math +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_t5_math` is a English model originally trained by DVLe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_t5_math_en_5.4.2_3.0_1724055065344.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_t5_math_en_5.4.2_3.0_1724055065344.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetuned_t5_math","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetuned_t5_math", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_t5_math| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/DVLe/finetuned_t5_math \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_pipeline_en.md new file mode 100644 index 00000000000000..4d6cca299c2d0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetuned_t5_math_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetuned_t5_math_pipeline pipeline T5Transformer from DVLe +author: John Snow Labs +name: finetuned_t5_math_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetuned_t5_math_pipeline` is a English model originally trained by DVLe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetuned_t5_math_pipeline_en_5.4.2_3.0_1724055110817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetuned_t5_math_pipeline_en_5.4.2_3.0_1724055110817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetuned_t5_math_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetuned_t5_math_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetuned_t5_math_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/DVLe/finetuned_t5_math + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_en.md new file mode 100644 index 00000000000000..643ca5b93db85e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetunedmt5_50k T5Transformer from baskotayunisha +author: John Snow Labs +name: finetunedmt5_50k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetunedmt5_50k` is a English model originally trained by baskotayunisha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetunedmt5_50k_en_5.4.2_3.0_1724094156256.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetunedmt5_50k_en_5.4.2_3.0_1724094156256.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetunedmt5_50k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetunedmt5_50k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetunedmt5_50k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/baskotayunisha/finetunedmt5-50k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_pipeline_en.md new file mode 100644 index 00000000000000..46605d3725d3c0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-finetunedmt5_50k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetunedmt5_50k_pipeline pipeline T5Transformer from baskotayunisha +author: John Snow Labs +name: finetunedmt5_50k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetunedmt5_50k_pipeline` is a English model originally trained by baskotayunisha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetunedmt5_50k_pipeline_en_5.4.2_3.0_1724094243204.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetunedmt5_50k_pipeline_en_5.4.2_3.0_1724094243204.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetunedmt5_50k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetunedmt5_50k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetunedmt5_50k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/baskotayunisha/finetunedmt5-50k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_en.md new file mode 100644 index 00000000000000..7a58e8e1cf1525 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_small_passage_evidence T5Transformer from erbacher +author: John Snow Labs +name: flan_small_passage_evidence +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_small_passage_evidence` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_small_passage_evidence_en_5.4.2_3.0_1724108036270.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_small_passage_evidence_en_5.4.2_3.0_1724108036270.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_small_passage_evidence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_small_passage_evidence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_small_passage_evidence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/erbacher/flan-small-passage-evidence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_pipeline_en.md new file mode 100644 index 00000000000000..054ad781c03fdc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_small_passage_evidence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_small_passage_evidence_pipeline pipeline T5Transformer from erbacher +author: John Snow Labs +name: flan_small_passage_evidence_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_small_passage_evidence_pipeline` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_small_passage_evidence_pipeline_en_5.4.2_3.0_1724108053196.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_small_passage_evidence_pipeline_en_5.4.2_3.0_1724108053196.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_small_passage_evidence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_small_passage_evidence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_small_passage_evidence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/erbacher/flan-small-passage-evidence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_en.md new file mode 100644 index 00000000000000..622c10a7855a23 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_1_6_xsum T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_6_xsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_6_xsum` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_xsum_en_5.4.2_3.0_1724073227077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_xsum_en_5.4.2_3.0_1724073227077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_1_6_xsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_1_6_xsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_6_xsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|754.0 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-6-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_pipeline_en.md new file mode 100644 index 00000000000000..6366443e2863a5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_1_6_xsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_1_6_xsum_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_6_xsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_6_xsum_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_xsum_pipeline_en_5.4.2_3.0_1724073269705.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_6_xsum_pipeline_en_5.4.2_3.0_1724073269705.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_1_6_xsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_1_6_xsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_6_xsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|754.0 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-6-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_en.md new file mode 100644 index 00000000000000..0faae15b2088ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_3_6_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_3_6_cnndm +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_3_6_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_cnndm_en_5.4.2_3.0_1724096745581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_cnndm_en_5.4.2_3.0_1724096745581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_3_6_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_3_6_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_3_6_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|860.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-3-6-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..27f8f06b538a43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_3_6_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_3_6_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_3_6_cnndm_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_3_6_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_cnndm_pipeline_en_5.4.2_3.0_1724096786821.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_cnndm_pipeline_en_5.4.2_3.0_1724096786821.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_3_6_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_3_6_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_3_6_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|860.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-3-6-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_en.md new file mode 100644 index 00000000000000..65da8e11c7c560 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_clara_med T5Transformer from CLARA-MeD +author: John Snow Labs +name: flan_t5_base_clara_med +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_clara_med` is a English model originally trained by CLARA-MeD. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_clara_med_en_5.4.2_3.0_1724111216420.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_clara_med_en_5.4.2_3.0_1724111216420.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_clara_med","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_clara_med", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_clara_med| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/CLARA-MeD/flan-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_pipeline_en.md new file mode 100644 index 00000000000000..69c21ef48d4187 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clara_med_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_clara_med_pipeline pipeline T5Transformer from CLARA-MeD +author: John Snow Labs +name: flan_t5_base_clara_med_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_clara_med_pipeline` is a English model originally trained by CLARA-MeD. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_clara_med_pipeline_en_5.4.2_3.0_1724111264546.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_clara_med_pipeline_en_5.4.2_3.0_1724111264546.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_clara_med_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_clara_med_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_clara_med_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/CLARA-MeD/flan-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_en.md new file mode 100644 index 00000000000000..9b12c8662963d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_clickbait_spoiling T5Transformer from Mike-HF +author: John Snow Labs +name: flan_t5_base_clickbait_spoiling +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_clickbait_spoiling` is a English model originally trained by Mike-HF. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_clickbait_spoiling_en_5.4.2_3.0_1724038897353.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_clickbait_spoiling_en_5.4.2_3.0_1724038897353.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_clickbait_spoiling","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_clickbait_spoiling", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_clickbait_spoiling| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Mike-HF/flan-t5-base-clickbait-spoiling \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_pipeline_en.md new file mode 100644 index 00000000000000..b9f14eb8c37053 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_clickbait_spoiling_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_clickbait_spoiling_pipeline pipeline T5Transformer from Mike-HF +author: John Snow Labs +name: flan_t5_base_clickbait_spoiling_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_clickbait_spoiling_pipeline` is a English model originally trained by Mike-HF. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_clickbait_spoiling_pipeline_en_5.4.2_3.0_1724038947366.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_clickbait_spoiling_pipeline_en_5.4.2_3.0_1724038947366.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_clickbait_spoiling_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_clickbait_spoiling_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_clickbait_spoiling_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Mike-HF/flan-t5-base-clickbait-spoiling + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_en.md new file mode 100644 index 00000000000000..49155f7e684fb5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_en_5.4.2_3.0_1724072005975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_en_5.4.2_3.0_1724072005975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-base-extraction-cnndm_8000-all-hint_precision-ep50 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline_en.md new file mode 100644 index 00000000000000..b3092fc45c8f66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline_en_5.4.2_3.0_1724072055187.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline_en_5.4.2_3.0_1724072055187.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extraction_cnndm_8000_all_hint_precision_ep50_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-base-extraction-cnndm_8000-all-hint_precision-ep50 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_en.md new file mode 100644 index 00000000000000..42d4a2d0e550d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_extraction_cnndm_8000_all_loss_ep50 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_base_extraction_cnndm_8000_all_loss_ep50 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extraction_cnndm_8000_all_loss_ep50` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_loss_ep50_en_5.4.2_3.0_1724077229592.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_loss_ep50_en_5.4.2_3.0_1724077229592.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_extraction_cnndm_8000_all_loss_ep50","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_extraction_cnndm_8000_all_loss_ep50", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extraction_cnndm_8000_all_loss_ep50| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-base-extraction-cnndm_8000-all-loss-ep50 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline_en.md new file mode 100644 index 00000000000000..ca37d8ed1d1c78 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline_en_5.4.2_3.0_1724077276397.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline_en_5.4.2_3.0_1724077276397.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_extraction_cnndm_8000_all_loss_ep50_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-base-extraction-cnndm_8000-all-loss-ep50 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_en.md new file mode 100644 index 00000000000000..fdf673173e675f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_adj_mts_bio_unique_dialogue T5Transformer from hankym +author: John Snow Labs +name: flan_t5_base_finetuned_adj_mts_bio_unique_dialogue +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_adj_mts_bio_unique_dialogue` is a English model originally trained by hankym. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_en_5.4.2_3.0_1724055325147.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_en_5.4.2_3.0_1724055325147.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_adj_mts_bio_unique_dialogue","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_adj_mts_bio_unique_dialogue", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_adj_mts_bio_unique_dialogue| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/hankym/flan_t5_base_finetuned_adj_MTS_bio_unique_dialogue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline_en.md new file mode 100644 index 00000000000000..631e847ab7cb54 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline pipeline T5Transformer from hankym +author: John Snow Labs +name: flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline` is a English model originally trained by hankym. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline_en_5.4.2_3.0_1724055497243.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline_en_5.4.2_3.0_1724055497243.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_adj_mts_bio_unique_dialogue_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/hankym/flan_t5_base_finetuned_adj_MTS_bio_unique_dialogue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_en.md new file mode 100644 index 00000000000000..b9396e035b7d58 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym T5Transformer from hankym +author: John Snow Labs +name: flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym` is a English model originally trained by hankym. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_en_5.4.2_3.0_1724037437079.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_en_5.4.2_3.0_1724037437079.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/hankym/flan_t5_base_finetuned_MTS_clinical_adj_unique_dialogue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline_en.md new file mode 100644 index 00000000000000..4b04799a4f3767 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline pipeline T5Transformer from hankym +author: John Snow Labs +name: flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline` is a English model originally trained by hankym. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline_en_5.4.2_3.0_1724037609212.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline_en_5.4.2_3.0_1724037609212.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_mts_clinical_adj_unique_dialogue_hankym_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/hankym/flan_t5_base_finetuned_MTS_clinical_adj_unique_dialogue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_en.md new file mode 100644 index 00000000000000..77422e285374bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem T5Transformer from agnesem +author: John Snow Labs +name: flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem` is a English model originally trained by agnesem. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_en_5.4.2_3.0_1724050867039.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_en_5.4.2_3.0_1724050867039.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/agnesem/flan_t5_base_finetuned_MTS_clinical_unique_dialogue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline_en.md new file mode 100644 index 00000000000000..4cba37151f2d1c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline pipeline T5Transformer from agnesem +author: John Snow Labs +name: flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline` is a English model originally trained by agnesem. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline_en_5.4.2_3.0_1724051045092.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline_en_5.4.2_3.0_1724051045092.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_finetuned_mts_clinical_unique_dialogue_agnesem_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.2 MB| + +## References + +https://huggingface.co/agnesem/flan_t5_base_finetuned_MTS_clinical_unique_dialogue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_en.md new file mode 100644 index 00000000000000..7a7fb4182b7739 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_fold_4 T5Transformer from research-dump +author: John Snow Labs +name: flan_t5_base_fold_4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_fold_4` is a English model originally trained by research-dump. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_fold_4_en_5.4.2_3.0_1724103317264.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_fold_4_en_5.4.2_3.0_1724103317264.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_fold_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_fold_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_fold_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-dump/flan-t5-base_fold_4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_pipeline_en.md new file mode 100644 index 00000000000000..154c90962f3627 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_fold_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_fold_4_pipeline pipeline T5Transformer from research-dump +author: John Snow Labs +name: flan_t5_base_fold_4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_fold_4_pipeline` is a English model originally trained by research-dump. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_fold_4_pipeline_en_5.4.2_3.0_1724103363766.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_fold_4_pipeline_en_5.4.2_3.0_1724103363766.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_fold_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_fold_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_fold_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-dump/flan-t5-base_fold_4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_en.md new file mode 100644 index 00000000000000..202dee4424b809 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_intent T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_intent +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_intent` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_intent_en_5.4.2_3.0_1724090023762.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_intent_en_5.4.2_3.0_1724090023762.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_intent","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_intent", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_intent| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-intent \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_pipeline_en.md new file mode 100644 index 00000000000000..98b140d43e5e1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_intent_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_intent_pipeline pipeline T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_intent_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_intent_pipeline` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_intent_pipeline_en_5.4.2_3.0_1724090069931.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_intent_pipeline_en_5.4.2_3.0_1724090069931.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_intent_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_intent_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_intent_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-intent + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_en.md new file mode 100644 index 00000000000000..6e19e6de1ff936 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_lumunot T5Transformer from lumunot +author: John Snow Labs +name: flan_t5_base_lumunot +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_lumunot` is a English model originally trained by lumunot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_lumunot_en_5.4.2_3.0_1724095964297.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_lumunot_en_5.4.2_3.0_1724095964297.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_lumunot","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_lumunot", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_lumunot| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lumunot/flan-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_pipeline_en.md new file mode 100644 index 00000000000000..d8a2ce4408af05 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_lumunot_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_lumunot_pipeline pipeline T5Transformer from lumunot +author: John Snow Labs +name: flan_t5_base_lumunot_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_lumunot_pipeline` is a English model originally trained by lumunot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_lumunot_pipeline_en_5.4.2_3.0_1724096010501.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_lumunot_pipeline_en_5.4.2_3.0_1724096010501.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_lumunot_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_lumunot_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_lumunot_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lumunot/flan-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_en.md new file mode 100644 index 00000000000000..cd3f2d8b1d0b9b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_qg_squad_5 T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_base_qg_squad_5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_qg_squad_5` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_5_en_5.4.2_3.0_1724062610088.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_5_en_5.4.2_3.0_1724062610088.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_qg_squad_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_qg_squad_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_qg_squad_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-base-qg-SQuAD-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_pipeline_en.md new file mode 100644 index 00000000000000..8202f2f4fe8ee3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_qg_squad_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_qg_squad_5_pipeline pipeline T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_base_qg_squad_5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_qg_squad_5_pipeline` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_5_pipeline_en_5.4.2_3.0_1724062657098.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_qg_squad_5_pipeline_en_5.4.2_3.0_1724062657098.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_qg_squad_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_qg_squad_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_qg_squad_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tarek23/flan-t5-base-qg-SQuAD-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_en.md new file mode 100644 index 00000000000000..8d42b145c30d57 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_rank_search T5Transformer from helliun +author: John Snow Labs +name: flan_t5_base_rank_search +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_rank_search` is a English model originally trained by helliun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_rank_search_en_5.4.2_3.0_1724095719100.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_rank_search_en_5.4.2_3.0_1724095719100.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_rank_search","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_rank_search", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_rank_search| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/helliun/flan-t5-base-rank-search \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_pipeline_en.md new file mode 100644 index 00000000000000..0ec99032ddbbea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_rank_search_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_rank_search_pipeline pipeline T5Transformer from helliun +author: John Snow Labs +name: flan_t5_base_rank_search_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_rank_search_pipeline` is a English model originally trained by helliun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_rank_search_pipeline_en_5.4.2_3.0_1724095766664.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_rank_search_pipeline_en_5.4.2_3.0_1724095766664.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_rank_search_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_rank_search_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_rank_search_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/helliun/flan-t5-base-rank-search + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_en.md new file mode 100644 index 00000000000000..fcaf129f2c3d7f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_tweet_qa T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_base_tweet_qa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_tweet_qa` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qa_en_5.4.2_3.0_1724078247023.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qa_en_5.4.2_3.0_1724078247023.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_tweet_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_tweet_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_tweet_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-base-tweet-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_pipeline_en.md new file mode 100644 index 00000000000000..c4504957a5f44b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_base_tweet_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_tweet_qa_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_base_tweet_qa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_tweet_qa_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qa_pipeline_en_5.4.2_3.0_1724078294425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_tweet_qa_pipeline_en_5.4.2_3.0_1724078294425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_tweet_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_tweet_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_tweet_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-base-tweet-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_en.md new file mode 100644 index 00000000000000..b66e7dd8319263 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_en_5.4.2_3.0_1724100118364.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_en_5.4.2_3.0_1724100118364.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep18-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..428597e2eb2e8e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline_en_5.4.2_3.0_1724100260089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline_en_5.4.2_3.0_1724100260089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep18_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep18-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_en.md new file mode 100644 index 00000000000000..811976b6508695 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_en_5.4.2_3.0_1724097366272.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_en_5.4.2_3.0_1724097366272.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep8-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..8d0e23c379914e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline_en_5.4.2_3.0_1724097519295.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline_en_5.4.2_3.0_1724097519295.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_400_ep8_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_400-ep8-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop_en.md new file mode 100644 index 00000000000000..a669a32e362b0d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop_en_5.4.2_3.0_1724102468370.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop_en_5.4.2_3.0_1724102468370.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_800_ep20_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_800-ep20-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_en.md new file mode 100644 index 00000000000000..396dc1d0122813 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_ep10 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_ep10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_ep10` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep10_en_5.4.2_3.0_1724074400057.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep10_en_5.4.2_3.0_1724074400057.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_ep10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_ep10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_ep10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-ep10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline_en.md new file mode 100644 index 00000000000000..34d2cae5e010ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline_en_5.4.2_3.0_1724074543734.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline_en_5.4.2_3.0_1724074543734.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_ep10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-ep10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_en.md new file mode 100644 index 00000000000000..15d07fccea59ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_loss_ep50 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_loss_ep50 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_loss_ep50` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_loss_ep50_en_5.4.2_3.0_1724109134330.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_loss_ep50_en_5.4.2_3.0_1724109134330.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_loss_ep50","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_400_loss_ep50", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_loss_ep50| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-loss-ep50 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline_en.md new file mode 100644 index 00000000000000..bb0c60a4238894 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline_en_5.4.2_3.0_1724109275084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline_en_5.4.2_3.0_1724109275084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_400_loss_ep50_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_400-loss-ep50 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_en.md new file mode 100644 index 00000000000000..b7367976e184bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_en_5.4.2_3.0_1724047721375.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_en_5.4.2_3.0_1724047721375.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_80-ep50-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..4084c144c9bc65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline_en_5.4.2_3.0_1724047856741.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline_en_5.4.2_3.0_1724047856741.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_80_ep50_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_80-ep50-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_en.md new file mode 100644 index 00000000000000..00d0f9c5ed6eb1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_extraction_all_cnn_8000_ep25_nonstop T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_all_cnn_8000_ep25_nonstop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_all_cnn_8000_ep25_nonstop` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_en_5.4.2_3.0_1724090229749.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_en_5.4.2_3.0_1724090229749.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_extraction_all_cnn_8000_ep25_nonstop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_extraction_all_cnn_8000_ep25_nonstop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_all_cnn_8000_ep25_nonstop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-all-cnn_8000-ep25-nonstop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline_en.md new file mode 100644 index 00000000000000..1b97afd38f1381 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline_en_5.4.2_3.0_1724090369941.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline_en_5.4.2_3.0_1724090369941.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_all_cnn_8000_ep25_nonstop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-all-cnn_8000-ep25-nonstop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_en.md new file mode 100644 index 00000000000000..467d9b41b9a9b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_extraction_cnndm_5000_all T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_cnndm_5000_all +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_cnndm_5000_all` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_5000_all_en_5.4.2_3.0_1724073592571.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_5000_all_en_5.4.2_3.0_1724073592571.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_extraction_cnndm_5000_all","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_extraction_cnndm_5000_all", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_cnndm_5000_all| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-cnndm_5000-all \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_pipeline_en.md new file mode 100644 index 00000000000000..e079d5fcaadcaf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_extraction_cnndm_5000_all_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_extraction_cnndm_5000_all_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_extraction_cnndm_5000_all_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_extraction_cnndm_5000_all_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_5000_all_pipeline_en_5.4.2_3.0_1724073740993.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_extraction_cnndm_5000_all_pipeline_en_5.4.2_3.0_1724073740993.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_extraction_cnndm_5000_all_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_extraction_cnndm_5000_all_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_extraction_cnndm_5000_all_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-extraction-cnndm_5000-all + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_en.md new file mode 100644 index 00000000000000..1ec95af94abd4b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_jonnylaw T5Transformer from Jonnylaw +author: John Snow Labs +name: flan_t5_large_jonnylaw +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_jonnylaw` is a English model originally trained by Jonnylaw. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_jonnylaw_en_5.4.2_3.0_1724104230094.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_jonnylaw_en_5.4.2_3.0_1724104230094.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_jonnylaw","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_jonnylaw", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_jonnylaw| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Jonnylaw/flan-t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_pipeline_en.md new file mode 100644 index 00000000000000..4ba03b06e88055 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_jonnylaw_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_jonnylaw_pipeline pipeline T5Transformer from Jonnylaw +author: John Snow Labs +name: flan_t5_large_jonnylaw_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_jonnylaw_pipeline` is a English model originally trained by Jonnylaw. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_jonnylaw_pipeline_en_5.4.2_3.0_1724104631813.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_jonnylaw_pipeline_en_5.4.2_3.0_1724104631813.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_jonnylaw_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_jonnylaw_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_jonnylaw_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Jonnylaw/flan-t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_en.md new file mode 100644 index 00000000000000..686ca9b9933838 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_medistill_28_hyperparameters_rougeave T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_28_hyperparameters_rougeave +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_28_hyperparameters_rougeave` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_rougeave_en_5.4.2_3.0_1724073272592.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_rougeave_en_5.4.2_3.0_1724073272592.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_hyperparameters_rougeave","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_medistill_28_hyperparameters_rougeave", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_28_hyperparameters_rougeave| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_28_hyperparameters_rougeave \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline_en.md new file mode 100644 index 00000000000000..e7f11c4d5ce411 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline pipeline T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline_en_5.4.2_3.0_1724073446358.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline_en_5.4.2_3.0_1724073446358.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_28_hyperparameters_rougeave_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_28_hyperparameters_rougeave + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_en.md new file mode 100644 index 00000000000000..49420464f6c5cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_multiqg T5Transformer from marksverdhei +author: John Snow Labs +name: flan_t5_large_multiqg +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_multiqg` is a English model originally trained by marksverdhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_multiqg_en_5.4.2_3.0_1724060769443.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_multiqg_en_5.4.2_3.0_1724060769443.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_multiqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_multiqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_multiqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/marksverdhei/flan-t5-large-multiqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_pipeline_en.md new file mode 100644 index 00000000000000..fd8b12f991fdd8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_large_multiqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_multiqg_pipeline pipeline T5Transformer from marksverdhei +author: John Snow Labs +name: flan_t5_large_multiqg_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_multiqg_pipeline` is a English model originally trained by marksverdhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_multiqg_pipeline_en_5.4.2_3.0_1724060903337.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_multiqg_pipeline_en_5.4.2_3.0_1724060903337.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_multiqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_multiqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_multiqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/marksverdhei/flan-t5-large-multiqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_en.md new file mode 100644 index 00000000000000..2261259e514995 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_qg_learningq_tarek_test_v2 T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_qg_learningq_tarek_test_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_qg_learningq_tarek_test_v2` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_qg_learningq_tarek_test_v2_en_5.4.2_3.0_1724100912823.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_qg_learningq_tarek_test_v2_en_5.4.2_3.0_1724100912823.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_qg_learningq_tarek_test_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_qg_learningq_tarek_test_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_qg_learningq_tarek_test_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/tarek23/flan-t5-qg-LearningQ-tarek-test-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_pipeline_en.md new file mode 100644 index 00000000000000..1a3df36a9572a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_qg_learningq_tarek_test_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_qg_learningq_tarek_test_v2_pipeline pipeline T5Transformer from tarek23 +author: John Snow Labs +name: flan_t5_qg_learningq_tarek_test_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_qg_learningq_tarek_test_v2_pipeline` is a English model originally trained by tarek23. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_qg_learningq_tarek_test_v2_pipeline_en_5.4.2_3.0_1724100929776.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_qg_learningq_tarek_test_v2_pipeline_en_5.4.2_3.0_1724100929776.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_qg_learningq_tarek_test_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_qg_learningq_tarek_test_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_qg_learningq_tarek_test_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/tarek23/flan-t5-qg-LearningQ-tarek-test-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_en.md new file mode 100644 index 00000000000000..f5ff79932466c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_3_6_xsum T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_3_6_xsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_3_6_xsum` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_3_6_xsum_en_5.4.2_3.0_1724073171670.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_3_6_xsum_en_5.4.2_3.0_1724073171670.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_3_6_xsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_3_6_xsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_3_6_xsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|314.4 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-3-6-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_pipeline_en.md new file mode 100644 index 00000000000000..d284d501f02dfe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_3_6_xsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_3_6_xsum_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_3_6_xsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_3_6_xsum_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_3_6_xsum_pipeline_en_5.4.2_3.0_1724073186652.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_3_6_xsum_pipeline_en_5.4.2_3.0_1724073186652.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_3_6_xsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_3_6_xsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_3_6_xsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|314.4 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-3-6-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_en.md new file mode 100644 index 00000000000000..67be95a63393c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_6_5_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_5_cnndm +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_5_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_5_cnndm_en_5.4.2_3.0_1724037760020.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_5_cnndm_en_5.4.2_3.0_1724037760020.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_6_5_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_6_5_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_5_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|326.2 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-5-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..da7c5cfcf9f7fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_6_5_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_6_5_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_small_6_5_cnndm_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_6_5_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_5_cnndm_pipeline_en_5.4.2_3.0_1724037778917.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_6_5_cnndm_pipeline_en_5.4.2_3.0_1724037778917.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_6_5_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_6_5_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_6_5_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|326.2 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-small-6-5-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_en.md new file mode 100644 index 00000000000000..04f00edb666323 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_analogy_permutation_domain T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_small_analogy_permutation_domain +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_analogy_permutation_domain` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_analogy_permutation_domain_en_5.4.2_3.0_1724097664311.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_analogy_permutation_domain_en_5.4.2_3.0_1724097664311.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_analogy_permutation_domain","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_analogy_permutation_domain", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_analogy_permutation_domain| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/research-backup/flan-t5-small-analogy-permutation-domain \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_pipeline_en.md new file mode 100644 index 00000000000000..700afc0349f136 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_analogy_permutation_domain_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_analogy_permutation_domain_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_small_analogy_permutation_domain_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_analogy_permutation_domain_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_analogy_permutation_domain_pipeline_en_5.4.2_3.0_1724097682828.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_analogy_permutation_domain_pipeline_en_5.4.2_3.0_1724097682828.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_analogy_permutation_domain_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_analogy_permutation_domain_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_analogy_permutation_domain_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/research-backup/flan-t5-small-analogy-permutation-domain + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_en.md new file mode 100644 index 00000000000000..89f06ce3211b6f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_asap_t5_f3_prompt_adherence T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t5_f3_prompt_adherence +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t5_f3_prompt_adherence` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t5_f3_prompt_adherence_en_5.4.2_3.0_1724109332297.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t5_f3_prompt_adherence_en_5.4.2_3.0_1724109332297.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_asap_t5_f3_prompt_adherence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_asap_t5_f3_prompt_adherence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t5_f3_prompt_adherence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t5_f3_prompt_adherence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_pipeline_en.md new file mode 100644 index 00000000000000..6ce3e47085d941 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_asap_t5_f3_prompt_adherence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_asap_t5_f3_prompt_adherence_pipeline pipeline T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t5_f3_prompt_adherence_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t5_f3_prompt_adherence_pipeline` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t5_f3_prompt_adherence_pipeline_en_5.4.2_3.0_1724109349541.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t5_f3_prompt_adherence_pipeline_en_5.4.2_3.0_1724109349541.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_asap_t5_f3_prompt_adherence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_asap_t5_f3_prompt_adherence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t5_f3_prompt_adherence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t5_f3_prompt_adherence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_en.md new file mode 100644 index 00000000000000..ab575aba8a98e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_botco_qa_finetuned_question_generation_context_only T5Transformer from IsaacBot +author: John Snow Labs +name: flan_t5_small_botco_qa_finetuned_question_generation_context_only +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_botco_qa_finetuned_question_generation_context_only` is a English model originally trained by IsaacBot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_botco_qa_finetuned_question_generation_context_only_en_5.4.2_3.0_1724027034950.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_botco_qa_finetuned_question_generation_context_only_en_5.4.2_3.0_1724027034950.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_botco_qa_finetuned_question_generation_context_only","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_botco_qa_finetuned_question_generation_context_only", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_botco_qa_finetuned_question_generation_context_only| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/IsaacBot/flan-t5-small-botco_QA-finetuned-question-generation-context-only \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline_en.md new file mode 100644 index 00000000000000..df730b1895d28d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline pipeline T5Transformer from IsaacBot +author: John Snow Labs +name: flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline` is a English model originally trained by IsaacBot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline_en_5.4.2_3.0_1724027051522.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline_en_5.4.2_3.0_1724027051522.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_botco_qa_finetuned_question_generation_context_only_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/IsaacBot/flan-t5-small-botco_QA-finetuned-question-generation-context-only + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_en.md new file mode 100644 index 00000000000000..d9dc9f2a472852 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_rl T5Transformer from vj1148 +author: John Snow Labs +name: flan_t5_small_rl +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_rl` is a English model originally trained by vj1148. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_rl_en_5.4.2_3.0_1724111299288.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_rl_en_5.4.2_3.0_1724111299288.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_rl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_rl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_rl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/vj1148/flan-t5-small-rl \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_pipeline_en.md new file mode 100644 index 00000000000000..c776ef4999dfe5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_rl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_rl_pipeline pipeline T5Transformer from vj1148 +author: John Snow Labs +name: flan_t5_small_rl_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_rl_pipeline` is a English model originally trained by vj1148. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_rl_pipeline_en_5.4.2_3.0_1724111316149.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_rl_pipeline_en_5.4.2_3.0_1724111316149.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_rl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_rl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_rl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/vj1148/flan-t5-small-rl + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_en.md new file mode 100644 index 00000000000000..e307e975a74032 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_samsum_konradszafer T5Transformer from KonradSzafer +author: John Snow Labs +name: flan_t5_small_samsum_konradszafer +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_konradszafer` is a English model originally trained by KonradSzafer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_konradszafer_en_5.4.2_3.0_1724070230215.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_konradszafer_en_5.4.2_3.0_1724070230215.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_samsum_konradszafer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_samsum_konradszafer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_konradszafer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/KonradSzafer/flan-t5-small-samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_pipeline_en.md new file mode 100644 index 00000000000000..edce758be72c83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_konradszafer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_samsum_konradszafer_pipeline pipeline T5Transformer from KonradSzafer +author: John Snow Labs +name: flan_t5_small_samsum_konradszafer_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_konradszafer_pipeline` is a English model originally trained by KonradSzafer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_konradszafer_pipeline_en_5.4.2_3.0_1724070247450.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_konradszafer_pipeline_en_5.4.2_3.0_1724070247450.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_samsum_konradszafer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_samsum_konradszafer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_konradszafer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/KonradSzafer/flan-t5-small-samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_en.md new file mode 100644 index 00000000000000..b8cdb9f847e7f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_samsum_pt_10 T5Transformer from PT-10 +author: John Snow Labs +name: flan_t5_small_samsum_pt_10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_pt_10` is a English model originally trained by PT-10. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_pt_10_en_5.4.2_3.0_1724083265101.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_pt_10_en_5.4.2_3.0_1724083265101.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_samsum_pt_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_samsum_pt_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_pt_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/PT-10/flan-t5-small-samsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_pipeline_en.md new file mode 100644 index 00000000000000..47743b629d198e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_samsum_pt_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_samsum_pt_10_pipeline pipeline T5Transformer from PT-10 +author: John Snow Labs +name: flan_t5_small_samsum_pt_10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_samsum_pt_10_pipeline` is a English model originally trained by PT-10. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_pt_10_pipeline_en_5.4.2_3.0_1724083283149.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_samsum_pt_10_pipeline_en_5.4.2_3.0_1724083283149.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_samsum_pt_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_samsum_pt_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_samsum_pt_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/PT-10/flan-t5-small-samsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_en.md new file mode 100644 index 00000000000000..daaf0bf00ab9e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_sft_small_hmq T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_small_sft_small_hmq +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_sft_small_hmq` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_sft_small_hmq_en_5.4.2_3.0_1724078658671.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_sft_small_hmq_en_5.4.2_3.0_1724078658671.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_sft_small_hmq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_sft_small_hmq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_sft_small_hmq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-small_sft_small_hmq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_pipeline_en.md new file mode 100644 index 00000000000000..9d307df3eeb83e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_sft_small_hmq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_sft_small_hmq_pipeline pipeline T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_small_sft_small_hmq_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_sft_small_hmq_pipeline` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_sft_small_hmq_pipeline_en_5.4.2_3.0_1724078718696.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_sft_small_hmq_pipeline_en_5.4.2_3.0_1724078718696.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_sft_small_hmq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_sft_small_hmq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_sft_small_hmq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-small_sft_small_hmq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_en.md new file mode 100644 index 00000000000000..39528ad18e7458 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_tweet_sentiment T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_sentiment +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_sentiment` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_sentiment_en_5.4.2_3.0_1724073626603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_sentiment_en_5.4.2_3.0_1724073626603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_tweet_sentiment","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_tweet_sentiment", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_sentiment| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-sentiment \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_pipeline_en.md new file mode 100644 index 00000000000000..f3d662a66d825c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_tweet_sentiment_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_tweet_sentiment_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_sentiment_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_sentiment_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_sentiment_pipeline_en_5.4.2_3.0_1724073644049.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_sentiment_pipeline_en_5.4.2_3.0_1724073644049.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_tweet_sentiment_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_tweet_sentiment_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_sentiment_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-sentiment + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_en.md new file mode 100644 index 00000000000000..f110b796cbccb9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_twitter_sentiment_analysis T5Transformer from thainq107 +author: John Snow Labs +name: flan_t5_small_twitter_sentiment_analysis +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_twitter_sentiment_analysis` is a English model originally trained by thainq107. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_twitter_sentiment_analysis_en_5.4.2_3.0_1724106239141.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_twitter_sentiment_analysis_en_5.4.2_3.0_1724106239141.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_twitter_sentiment_analysis","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_twitter_sentiment_analysis", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_twitter_sentiment_analysis| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/thainq107/flan-t5-small-twitter-sentiment-analysis \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_pipeline_en.md new file mode 100644 index 00000000000000..ee3fb8cd07880a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_small_twitter_sentiment_analysis_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_twitter_sentiment_analysis_pipeline pipeline T5Transformer from thainq107 +author: John Snow Labs +name: flan_t5_small_twitter_sentiment_analysis_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_twitter_sentiment_analysis_pipeline` is a English model originally trained by thainq107. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_twitter_sentiment_analysis_pipeline_en_5.4.2_3.0_1724106258008.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_twitter_sentiment_analysis_pipeline_en_5.4.2_3.0_1724106258008.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_twitter_sentiment_analysis_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_twitter_sentiment_analysis_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_twitter_sentiment_analysis_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/thainq107/flan-t5-small-twitter-sentiment-analysis + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_en.md new file mode 100644 index 00000000000000..94fd0393ad8efb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_xlsum T5Transformer from SadiulArefin +author: John Snow Labs +name: flan_t5_xlsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_xlsum` is a English model originally trained by SadiulArefin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_xlsum_en_5.4.2_3.0_1724034627975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_xlsum_en_5.4.2_3.0_1724034627975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_xlsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_xlsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_xlsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SadiulArefin/flan-t5-xlsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_pipeline_en.md new file mode 100644 index 00000000000000..71b0245facc82a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flan_t5_xlsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_xlsum_pipeline pipeline T5Transformer from SadiulArefin +author: John Snow Labs +name: flan_t5_xlsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_xlsum_pipeline` is a English model originally trained by SadiulArefin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_xlsum_pipeline_en_5.4.2_3.0_1724034675748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_xlsum_pipeline_en_5.4.2_3.0_1724034675748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_xlsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_xlsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_xlsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/SadiulArefin/flan-t5-xlsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_en.md new file mode 100644 index 00000000000000..c07c8f6684f190 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_mwpgen T5Transformer from apoorv627g +author: John Snow Labs +name: flant5_mwpgen +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_mwpgen` is a English model originally trained by apoorv627g. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_mwpgen_en_5.4.2_3.0_1724055955929.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_mwpgen_en_5.4.2_3.0_1724055955929.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_mwpgen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_mwpgen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_mwpgen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/apoorv627g/FlanT5_MWPgen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_pipeline_en.md new file mode 100644 index 00000000000000..babd43fb439d61 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_mwpgen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_mwpgen_pipeline pipeline T5Transformer from apoorv627g +author: John Snow Labs +name: flant5_mwpgen_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_mwpgen_pipeline` is a English model originally trained by apoorv627g. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_mwpgen_pipeline_en_5.4.2_3.0_1724056001939.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_mwpgen_pipeline_en_5.4.2_3.0_1724056001939.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_mwpgen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_mwpgen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_mwpgen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/apoorv627g/FlanT5_MWPgen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_en.md new file mode 100644 index 00000000000000..ffc18b535d082d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_sanar085 T5Transformer from sanar085 +author: John Snow Labs +name: flant5_sanar085 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_sanar085` is a English model originally trained by sanar085. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_sanar085_en_5.4.2_3.0_1724068324778.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_sanar085_en_5.4.2_3.0_1724068324778.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_sanar085","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_sanar085", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_sanar085| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sanar085/flant5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_pipeline_en.md new file mode 100644 index 00000000000000..2015ee19ae6e63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_sanar085_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_sanar085_pipeline pipeline T5Transformer from sanar085 +author: John Snow Labs +name: flant5_sanar085_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_sanar085_pipeline` is a English model originally trained by sanar085. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_sanar085_pipeline_en_5.4.2_3.0_1724068370941.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_sanar085_pipeline_en_5.4.2_3.0_1724068370941.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_sanar085_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_sanar085_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_sanar085_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sanar085/flant5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_en.md new file mode 100644 index 00000000000000..409ae5450c67da --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_small_finetuning_v2 T5Transformer from tuquyennnn +author: John Snow Labs +name: flant5_small_finetuning_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_small_finetuning_v2` is a English model originally trained by tuquyennnn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_small_finetuning_v2_en_5.4.2_3.0_1724099120017.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_small_finetuning_v2_en_5.4.2_3.0_1724099120017.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_small_finetuning_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_small_finetuning_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_small_finetuning_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|853.2 KB| + +## References + +https://huggingface.co/tuquyennnn/flant5-small-finetuning-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_pipeline_en.md new file mode 100644 index 00000000000000..920df5c6ee84a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_small_finetuning_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_small_finetuning_v2_pipeline pipeline T5Transformer from tuquyennnn +author: John Snow Labs +name: flant5_small_finetuning_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_small_finetuning_v2_pipeline` is a English model originally trained by tuquyennnn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_small_finetuning_v2_pipeline_en_5.4.2_3.0_1724099121399.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_small_finetuning_v2_pipeline_en_5.4.2_3.0_1724099121399.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_small_finetuning_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_small_finetuning_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_small_finetuning_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|856.3 KB| + +## References + +https://huggingface.co/tuquyennnn/flant5-small-finetuning-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_en.md new file mode 100644 index 00000000000000..6f26b9b20bda5f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5_summarymodel T5Transformer from mohammedRiad +author: John Snow Labs +name: flant5_summarymodel +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_summarymodel` is a English model originally trained by mohammedRiad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_summarymodel_en_5.4.2_3.0_1724067035154.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_summarymodel_en_5.4.2_3.0_1724067035154.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5_summarymodel","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5_summarymodel", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_summarymodel| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mohammedRiad/flant5-summaryModel \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_pipeline_en.md new file mode 100644 index 00000000000000..835d801e1e90c0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5_summarymodel_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5_summarymodel_pipeline pipeline T5Transformer from mohammedRiad +author: John Snow Labs +name: flant5_summarymodel_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5_summarymodel_pipeline` is a English model originally trained by mohammedRiad. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5_summarymodel_pipeline_en_5.4.2_3.0_1724067085561.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5_summarymodel_pipeline_en_5.4.2_3.0_1724067085561.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5_summarymodel_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5_summarymodel_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5_summarymodel_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mohammedRiad/flant5-summaryModel + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_en.md new file mode 100644 index 00000000000000..17a3adb8ff5a63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flant5s_jp10000 T5Transformer from mimi33 +author: John Snow Labs +name: flant5s_jp10000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5s_jp10000` is a English model originally trained by mimi33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5s_jp10000_en_5.4.2_3.0_1724040500919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5s_jp10000_en_5.4.2_3.0_1724040500919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flant5s_jp10000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flant5s_jp10000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5s_jp10000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/mimi33/flant5s-JP10000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_pipeline_en.md new file mode 100644 index 00000000000000..ebae0a4e503280 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-flant5s_jp10000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flant5s_jp10000_pipeline pipeline T5Transformer from mimi33 +author: John Snow Labs +name: flant5s_jp10000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flant5s_jp10000_pipeline` is a English model originally trained by mimi33. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flant5s_jp10000_pipeline_en_5.4.2_3.0_1724040516739.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flant5s_jp10000_pipeline_en_5.4.2_3.0_1724040516739.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flant5s_jp10000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flant5s_jp10000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flant5s_jp10000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/mimi33/flant5s-JP10000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_en.md b/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_en.md new file mode 100644 index 00000000000000..bc3d3d4932311b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_aviationqa_aviationcorpus_20 T5Transformer from sakharamg +author: John Snow Labs +name: ft_aviationqa_aviationcorpus_20 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aviationqa_aviationcorpus_20` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aviationqa_aviationcorpus_20_en_5.4.2_3.0_1724075945169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aviationqa_aviationcorpus_20_en_5.4.2_3.0_1724075945169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_aviationqa_aviationcorpus_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_aviationqa_aviationcorpus_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aviationqa_aviationcorpus_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aviationqa_aviationcorpus_20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_pipeline_en.md new file mode 100644 index 00000000000000..3c74a8f82637f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ft_aviationqa_aviationcorpus_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ft_aviationqa_aviationcorpus_20_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: ft_aviationqa_aviationcorpus_20_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aviationqa_aviationcorpus_20_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aviationqa_aviationcorpus_20_pipeline_en_5.4.2_3.0_1724076086581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aviationqa_aviationcorpus_20_pipeline_en_5.4.2_3.0_1724076086581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ft_aviationqa_aviationcorpus_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ft_aviationqa_aviationcorpus_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aviationqa_aviationcorpus_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aviationqa_aviationcorpus_20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_en.md b/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_en.md new file mode 100644 index 00000000000000..881f1d635b58d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English german4ep_4b T5Transformer from Bistolero +author: John Snow Labs +name: german4ep_4b +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`german4ep_4b` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/german4ep_4b_en_5.4.2_3.0_1724089348491.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/german4ep_4b_en_5.4.2_3.0_1724089348491.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("german4ep_4b","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("german4ep_4b", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|german4ep_4b| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Bistolero/german4ep_4b \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_pipeline_en.md new file mode 100644 index 00000000000000..c488e0695f6577 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-german4ep_4b_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English german4ep_4b_pipeline pipeline T5Transformer from Bistolero +author: John Snow Labs +name: german4ep_4b_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`german4ep_4b_pipeline` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/german4ep_4b_pipeline_en_5.4.2_3.0_1724089518393.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/german4ep_4b_pipeline_en_5.4.2_3.0_1724089518393.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("german4ep_4b_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("german4ep_4b_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|german4ep_4b_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Bistolero/german4ep_4b + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_en.md b/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_en.md new file mode 100644 index 00000000000000..f2601ef825956d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English german_40k_final T5Transformer from Bistolero +author: John Snow Labs +name: german_40k_final +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`german_40k_final` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/german_40k_final_en_5.4.2_3.0_1724050717844.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/german_40k_final_en_5.4.2_3.0_1724050717844.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("german_40k_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("german_40k_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|german_40k_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/german_40k_final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_pipeline_en.md new file mode 100644 index 00000000000000..e7037204adc6ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-german_40k_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English german_40k_final_pipeline pipeline T5Transformer from Bistolero +author: John Snow Labs +name: german_40k_final_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`german_40k_final_pipeline` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/german_40k_final_pipeline_en_5.4.2_3.0_1724050879112.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/german_40k_final_pipeline_en_5.4.2_3.0_1724050879112.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("german_40k_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("german_40k_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|german_40k_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/german_40k_final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_en.md b/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_en.md new file mode 100644 index 00000000000000..e6c26d151b3593 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English google_flan_t5_large_alpaca T5Transformer from reasonwang +author: John Snow Labs +name: google_flan_t5_large_alpaca +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_flan_t5_large_alpaca` is a English model originally trained by reasonwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_flan_t5_large_alpaca_en_5.4.2_3.0_1724041814506.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_flan_t5_large_alpaca_en_5.4.2_3.0_1724041814506.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("google_flan_t5_large_alpaca","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("google_flan_t5_large_alpaca", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_flan_t5_large_alpaca| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/reasonwang/google-flan-t5-large-alpaca \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_pipeline_en.md new file mode 100644 index 00000000000000..a2031cbd8baa82 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-google_flan_t5_large_alpaca_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English google_flan_t5_large_alpaca_pipeline pipeline T5Transformer from reasonwang +author: John Snow Labs +name: google_flan_t5_large_alpaca_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_flan_t5_large_alpaca_pipeline` is a English model originally trained by reasonwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_flan_t5_large_alpaca_pipeline_en_5.4.2_3.0_1724041953386.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_flan_t5_large_alpaca_pipeline_en_5.4.2_3.0_1724041953386.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("google_flan_t5_large_alpaca_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("google_flan_t5_large_alpaca_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_flan_t5_large_alpaca_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/reasonwang/google-flan-t5-large-alpaca + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_en.md b/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_en.md new file mode 100644 index 00000000000000..4a9c3b8543b5b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English google_t5_for_summary T5Transformer from MayankMoon +author: John Snow Labs +name: google_t5_for_summary +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_t5_for_summary` is a English model originally trained by MayankMoon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_t5_for_summary_en_5.4.2_3.0_1724062678731.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_t5_for_summary_en_5.4.2_3.0_1724062678731.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("google_t5_for_summary","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("google_t5_for_summary", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_t5_for_summary| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|330.7 MB| + +## References + +https://huggingface.co/MayankMoon/google-t5-for-summary \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_pipeline_en.md new file mode 100644 index 00000000000000..94e88b633ca9e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-google_t5_for_summary_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English google_t5_for_summary_pipeline pipeline T5Transformer from MayankMoon +author: John Snow Labs +name: google_t5_for_summary_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`google_t5_for_summary_pipeline` is a English model originally trained by MayankMoon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/google_t5_for_summary_pipeline_en_5.4.2_3.0_1724062702045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/google_t5_for_summary_pipeline_en_5.4.2_3.0_1724062702045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("google_t5_for_summary_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("google_t5_for_summary_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|google_t5_for_summary_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|330.7 MB| + +## References + +https://huggingface.co/MayankMoon/google-t5-for-summary + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_en.md b/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_en.md new file mode 100644 index 00000000000000..afe3a75470b5e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_ep_v3 T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_ep_v3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_ep_v3` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v3_en_5.4.2_3.0_1724103036778.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v3_en_5.4.2_3.0_1724103036778.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_ep_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_ep_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_ep_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-ep-v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_pipeline_en.md new file mode 100644 index 00000000000000..3fdadf82907d8f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-happy_transformer_t5_base_grammar_correction_ep_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_ep_v3_pipeline pipeline T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_ep_v3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_ep_v3_pipeline` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v3_pipeline_en_5.4.2_3.0_1724103084935.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_ep_v3_pipeline_en_5.4.2_3.0_1724103084935.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("happy_transformer_t5_base_grammar_correction_ep_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("happy_transformer_t5_base_grammar_correction_ep_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_ep_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-ep-v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_en.md new file mode 100644 index 00000000000000..8c1248fcd8b3d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English hate_speech_detection_vit5_base T5Transformer from baohl00 +author: John Snow Labs +name: hate_speech_detection_vit5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`hate_speech_detection_vit5_base` is a English model originally trained by baohl00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/hate_speech_detection_vit5_base_en_5.4.2_3.0_1724069607258.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/hate_speech_detection_vit5_base_en_5.4.2_3.0_1724069607258.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("hate_speech_detection_vit5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("hate_speech_detection_vit5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|hate_speech_detection_vit5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/baohl00/hate-speech-detection-vit5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_pipeline_en.md new file mode 100644 index 00000000000000..98e6286e80d74f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-hate_speech_detection_vit5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English hate_speech_detection_vit5_base_pipeline pipeline T5Transformer from baohl00 +author: John Snow Labs +name: hate_speech_detection_vit5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`hate_speech_detection_vit5_base_pipeline` is a English model originally trained by baohl00. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/hate_speech_detection_vit5_base_pipeline_en_5.4.2_3.0_1724069658786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/hate_speech_detection_vit5_base_pipeline_en_5.4.2_3.0_1724069658786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("hate_speech_detection_vit5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("hate_speech_detection_vit5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|hate_speech_detection_vit5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/baohl00/hate-speech-detection-vit5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-het5_base_he.md b/docs/_posts/ahmedlone127/2024-08-19-het5_base_he.md new file mode 100644 index 00000000000000..d9c90b6d1d70f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-het5_base_he.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Hebrew het5_base T5Transformer from imvladikon +author: John Snow Labs +name: het5_base +date: 2024-08-19 +tags: [he, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: he +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`het5_base` is a Hebrew model originally trained by imvladikon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/het5_base_he_5.4.2_3.0_1724036899614.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/het5_base_he_5.4.2_3.0_1724036899614.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("het5_base","he") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("het5_base", "he") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|het5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|he| +|Size:|511.6 MB| + +## References + +https://huggingface.co/imvladikon/het5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-het5_base_pipeline_he.md b/docs/_posts/ahmedlone127/2024-08-19-het5_base_pipeline_he.md new file mode 100644 index 00000000000000..5a9221a27d4831 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-het5_base_pipeline_he.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Hebrew het5_base_pipeline pipeline T5Transformer from imvladikon +author: John Snow Labs +name: het5_base_pipeline +date: 2024-08-19 +tags: [he, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: he +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`het5_base_pipeline` is a Hebrew model originally trained by imvladikon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/het5_base_pipeline_he_5.4.2_3.0_1724037064457.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/het5_base_pipeline_he_5.4.2_3.0_1724037064457.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("het5_base_pipeline", lang = "he") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("het5_base_pipeline", lang = "he") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|het5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|he| +|Size:|511.6 MB| + +## References + +https://huggingface.co/imvladikon/het5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-indo_t5_base_nusax_en.md b/docs/_posts/ahmedlone127/2024-08-19-indo_t5_base_nusax_en.md new file mode 100644 index 00000000000000..d0731fec8ee597 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-indo_t5_base_nusax_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English indo_t5_base_nusax T5Transformer from LazarusNLP +author: John Snow Labs +name: indo_t5_base_nusax +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`indo_t5_base_nusax` is a English model originally trained by LazarusNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/indo_t5_base_nusax_en_5.4.2_3.0_1724063500842.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/indo_t5_base_nusax_en_5.4.2_3.0_1724063500842.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("indo_t5_base_nusax","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("indo_t5_base_nusax", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|indo_t5_base_nusax| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/LazarusNLP/indo-t5-base-nusax \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_en.md b/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_en.md new file mode 100644 index 00000000000000..92f7f332b875f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English indonesiant5_gec T5Transformer from treshnanda +author: John Snow Labs +name: indonesiant5_gec +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`indonesiant5_gec` is a English model originally trained by treshnanda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/indonesiant5_gec_en_5.4.2_3.0_1724084699624.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/indonesiant5_gec_en_5.4.2_3.0_1724084699624.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("indonesiant5_gec","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("indonesiant5_gec", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|indonesiant5_gec| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|520.7 MB| + +## References + +https://huggingface.co/treshnanda/IndonesianT5-GEC \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_pipeline_en.md new file mode 100644 index 00000000000000..03900dbe2007df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-indonesiant5_gec_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English indonesiant5_gec_pipeline pipeline T5Transformer from treshnanda +author: John Snow Labs +name: indonesiant5_gec_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`indonesiant5_gec_pipeline` is a English model originally trained by treshnanda. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/indonesiant5_gec_pipeline_en_5.4.2_3.0_1724084874926.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/indonesiant5_gec_pipeline_en_5.4.2_3.0_1724084874926.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("indonesiant5_gec_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("indonesiant5_gec_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|indonesiant5_gec_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|520.7 MB| + +## References + +https://huggingface.co/treshnanda/IndonesianT5-GEC + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_en.md b/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_en.md new file mode 100644 index 00000000000000..751d43b313fbbc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English interview_t5_base_sft T5Transformer from celise88 +author: John Snow Labs +name: interview_t5_base_sft +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`interview_t5_base_sft` is a English model originally trained by celise88. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/interview_t5_base_sft_en_5.4.2_3.0_1724039892234.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/interview_t5_base_sft_en_5.4.2_3.0_1724039892234.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("interview_t5_base_sft","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("interview_t5_base_sft", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|interview_t5_base_sft| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/celise88/interview-t5-base-sft \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_pipeline_en.md new file mode 100644 index 00000000000000..b091d08415b949 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-interview_t5_base_sft_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English interview_t5_base_sft_pipeline pipeline T5Transformer from celise88 +author: John Snow Labs +name: interview_t5_base_sft_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`interview_t5_base_sft_pipeline` is a English model originally trained by celise88. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/interview_t5_base_sft_pipeline_en_5.4.2_3.0_1724039937956.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/interview_t5_base_sft_pipeline_en_5.4.2_3.0_1724039937956.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("interview_t5_base_sft_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("interview_t5_base_sft_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|interview_t5_base_sft_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/celise88/interview-t5-base-sft + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_en.md b/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_en.md new file mode 100644 index 00000000000000..499a8fea8cc3b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kaggle_freezed_english_german T5Transformer from br0hum +author: John Snow Labs +name: kaggle_freezed_english_german +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kaggle_freezed_english_german` is a English model originally trained by br0hum. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kaggle_freezed_english_german_en_5.4.2_3.0_1724033805836.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kaggle_freezed_english_german_en_5.4.2_3.0_1724033805836.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kaggle_freezed_english_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kaggle_freezed_english_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kaggle_freezed_english_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/br0hum/Kaggle-freezed-en-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_pipeline_en.md new file mode 100644 index 00000000000000..469aa8230a2983 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kaggle_freezed_english_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kaggle_freezed_english_german_pipeline pipeline T5Transformer from br0hum +author: John Snow Labs +name: kaggle_freezed_english_german_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kaggle_freezed_english_german_pipeline` is a English model originally trained by br0hum. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kaggle_freezed_english_german_pipeline_en_5.4.2_3.0_1724033854669.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kaggle_freezed_english_german_pipeline_en_5.4.2_3.0_1724033854669.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kaggle_freezed_english_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kaggle_freezed_english_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kaggle_freezed_english_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/br0hum/Kaggle-freezed-en-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_en.md b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_en.md new file mode 100644 index 00000000000000..19fd03928e69fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ke_t5_base_bongsoo_english_korean T5Transformer from chunwoolee0 +author: John Snow Labs +name: ke_t5_base_bongsoo_english_korean +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_base_bongsoo_english_korean` is a English model originally trained by chunwoolee0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_base_bongsoo_english_korean_en_5.4.2_3.0_1724042953919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_base_bongsoo_english_korean_en_5.4.2_3.0_1724042953919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ke_t5_base_bongsoo_english_korean","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ke_t5_base_bongsoo_english_korean", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_base_bongsoo_english_korean| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/chunwoolee0/ke_t5_base_bongsoo_en_ko \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_pipeline_en.md new file mode 100644 index 00000000000000..ca52ef0891a155 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_bongsoo_english_korean_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ke_t5_base_bongsoo_english_korean_pipeline pipeline T5Transformer from chunwoolee0 +author: John Snow Labs +name: ke_t5_base_bongsoo_english_korean_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_base_bongsoo_english_korean_pipeline` is a English model originally trained by chunwoolee0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_base_bongsoo_english_korean_pipeline_en_5.4.2_3.0_1724043018569.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_base_bongsoo_english_korean_pipeline_en_5.4.2_3.0_1724043018569.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ke_t5_base_bongsoo_english_korean_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ke_t5_base_bongsoo_english_korean_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_base_bongsoo_english_korean_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/chunwoolee0/ke_t5_base_bongsoo_en_ko + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_en.md b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_en.md new file mode 100644 index 00000000000000..f5132ab4cec723 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ke_t5_base_korean_aihub_paper_summary T5Transformer from mimi +author: John Snow Labs +name: ke_t5_base_korean_aihub_paper_summary +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_base_korean_aihub_paper_summary` is a English model originally trained by mimi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_base_korean_aihub_paper_summary_en_5.4.2_3.0_1724087581495.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_base_korean_aihub_paper_summary_en_5.4.2_3.0_1724087581495.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ke_t5_base_korean_aihub_paper_summary","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ke_t5_base_korean_aihub_paper_summary", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_base_korean_aihub_paper_summary| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/mimi/ke-t5-base-ko-AIHub-paper-summary \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_pipeline_en.md new file mode 100644 index 00000000000000..891e915e68f718 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ke_t5_base_korean_aihub_paper_summary_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ke_t5_base_korean_aihub_paper_summary_pipeline pipeline T5Transformer from mimi +author: John Snow Labs +name: ke_t5_base_korean_aihub_paper_summary_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ke_t5_base_korean_aihub_paper_summary_pipeline` is a English model originally trained by mimi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ke_t5_base_korean_aihub_paper_summary_pipeline_en_5.4.2_3.0_1724087642620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ke_t5_base_korean_aihub_paper_summary_pipeline_en_5.4.2_3.0_1724087642620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ke_t5_base_korean_aihub_paper_summary_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ke_t5_base_korean_aihub_paper_summary_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ke_t5_base_korean_aihub_paper_summary_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/mimi/ke-t5-base-ko-AIHub-paper-summary + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_en.md new file mode 100644 index 00000000000000..965be80c53973d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English keyword_generator_2 T5Transformer from lucazed +author: John Snow Labs +name: keyword_generator_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`keyword_generator_2` is a English model originally trained by lucazed. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/keyword_generator_2_en_5.4.2_3.0_1724068530665.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/keyword_generator_2_en_5.4.2_3.0_1724068530665.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("keyword_generator_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("keyword_generator_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|keyword_generator_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lucazed/keyword-generator-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_pipeline_en.md new file mode 100644 index 00000000000000..b1d3879038fd32 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-keyword_generator_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English keyword_generator_2_pipeline pipeline T5Transformer from lucazed +author: John Snow Labs +name: keyword_generator_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`keyword_generator_2_pipeline` is a English model originally trained by lucazed. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/keyword_generator_2_pipeline_en_5.4.2_3.0_1724068578093.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/keyword_generator_2_pipeline_en_5.4.2_3.0_1724068578093.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("keyword_generator_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("keyword_generator_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|keyword_generator_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lucazed/keyword-generator-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_en.md new file mode 100644 index 00000000000000..d9f2811d410b30 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_asopl T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_asopl +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_asopl` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_en_5.4.2_3.0_1724080043374.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_en_5.4.2_3.0_1724080043374.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_asopl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_asopl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_asopl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_ASOPL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_en.md new file mode 100644 index 00000000000000..ab29a88d88814a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_asopl_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_asopl_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_asopl_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_v2_en_5.4.2_3.0_1724076508495.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_v2_en_5.4.2_3.0_1724076508495.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_asopl_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_asopl_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_asopl_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_ASOPL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_pipeline_en.md new file mode 100644 index 00000000000000..2d0cf983673169 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_asopl_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_asopl_v2_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_asopl_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_asopl_v2_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_v2_pipeline_en_5.4.2_3.0_1724076678440.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_asopl_v2_pipeline_en_5.4.2_3.0_1724076678440.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_asopl_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_asopl_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_asopl_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_ASOPL_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_en.md new file mode 100644 index 00000000000000..8417202a85468c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_oapsl_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_oapsl_v5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_oapsl_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_oapsl_v5_en_5.4.2_3.0_1724081322153.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_oapsl_v5_en_5.4.2_3.0_1724081322153.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_oapsl_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_oapsl_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_oapsl_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OAPSL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_pipeline_en.md new file mode 100644 index 00000000000000..93cfe0c290acda --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_oapsl_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_oapsl_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_oapsl_v5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_oapsl_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_oapsl_v5_pipeline_en_5.4.2_3.0_1724081489786.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_oapsl_v5_pipeline_en_5.4.2_3.0_1724081489786.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_oapsl_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_oapsl_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_oapsl_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OAPSL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_en.md new file mode 100644 index 00000000000000..c9215d7846ca69 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_opasl T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_opasl +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_opasl` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_opasl_en_5.4.2_3.0_1724071315294.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_opasl_en_5.4.2_3.0_1724071315294.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_opasl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_opasl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_opasl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OPASL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_pipeline_en.md new file mode 100644 index 00000000000000..b2c095144cf34b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_opasl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_opasl_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_opasl_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_opasl_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_opasl_pipeline_en_5.4.2_3.0_1724071503169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_opasl_pipeline_en_5.4.2_3.0_1724071503169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_opasl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_opasl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_opasl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_OPASL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_en.md new file mode 100644 index 00000000000000..b2ec367d65cef5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_psoal_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_psoal_v5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_psoal_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_psoal_v5_en_5.4.2_3.0_1724059769097.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_psoal_v5_en_5.4.2_3.0_1724059769097.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_psoal_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_psoal_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_psoal_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_PSOAL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_pipeline_en.md new file mode 100644 index 00000000000000..c6ffedb269ae36 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_psoal_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_psoal_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_psoal_v5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_psoal_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_psoal_v5_pipeline_en_5.4.2_3.0_1724059952429.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_psoal_v5_pipeline_en_5.4.2_3.0_1724059952429.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_psoal_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_psoal_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_psoal_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_PSOAL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_en.md new file mode 100644 index 00000000000000..c936670ff301b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_soapl_v6 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_soapl_v6 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_soapl_v6` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_soapl_v6_en_5.4.2_3.0_1724072737635.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_soapl_v6_en_5.4.2_3.0_1724072737635.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_soapl_v6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_soapl_v6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_soapl_v6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SOAPL_v6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_pipeline_en.md new file mode 100644 index 00000000000000..2c401a780d03af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_soapl_v6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_soapl_v6_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_soapl_v6_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_soapl_v6_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_soapl_v6_pipeline_en_5.4.2_3.0_1724072919772.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_soapl_v6_pipeline_en_5.4.2_3.0_1724072919772.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_soapl_v6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_soapl_v6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_soapl_v6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SOAPL_v6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_en.md new file mode 100644 index 00000000000000..641420f8531581 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_sopal_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_sopal_v5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_sopal_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sopal_v5_en_5.4.2_3.0_1724045201033.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sopal_v5_en_5.4.2_3.0_1724045201033.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_sopal_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_sopal_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_sopal_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SOPAL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_pipeline_en.md new file mode 100644 index 00000000000000..e8d6ce5237d735 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_sopal_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_sopal_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_sopal_v5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_sopal_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sopal_v5_pipeline_en_5.4.2_3.0_1724045394538.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_sopal_v5_pipeline_en_5.4.2_3.0_1724045394538.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_sopal_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_sopal_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_sopal_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_SOPAL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_en.md new file mode 100644 index 00000000000000..08471a0ed42e99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_asopl_v4 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_asopl_v4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_asopl_v4` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_asopl_v4_en_5.4.2_3.0_1724110027269.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_asopl_v4_en_5.4.2_3.0_1724110027269.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_asopl_v4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_asopl_v4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_asopl_v4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_ASOPL_v4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_pipeline_en.md new file mode 100644 index 00000000000000..53dca74032d795 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_asopl_v4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_asopl_v4_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_asopl_v4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_asopl_v4_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_asopl_v4_pipeline_en_5.4.2_3.0_1724110203338.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_asopl_v4_pipeline_en_5.4.2_3.0_1724110203338.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_asopl_v4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_asopl_v4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_asopl_v4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_ASOPL_v4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_en.md new file mode 100644 index 00000000000000..1b77b49713bbcf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_pasol_v3 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_pasol_v3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_pasol_v3` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_pasol_v3_en_5.4.2_3.0_1724099782438.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_pasol_v3_en_5.4.2_3.0_1724099782438.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_pasol_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_pasol_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_pasol_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_PASOL_v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_pipeline_en.md new file mode 100644 index 00000000000000..1fcba9cb9d5b28 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_pasol_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_pasol_v3_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_pasol_v3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_pasol_v3_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_pasol_v3_pipeline_en_5.4.2_3.0_1724099966031.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_pasol_v3_pipeline_en_5.4.2_3.0_1724099966031.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_pasol_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_pasol_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_pasol_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_PASOL_v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_poasl_v5_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_poasl_v5_en.md new file mode 100644 index 00000000000000..a6520c375b5c65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_poasl_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_poasl_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_poasl_v5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_poasl_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v5_en_5.4.2_3.0_1724102650646.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_poasl_v5_en_5.4.2_3.0_1724102650646.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_poasl_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_poasl_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_poasl_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POASL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v2_en.md new file mode 100644 index 00000000000000..1ffbc479843b4e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_posal_v2 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_posal_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_posal_v2` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v2_en_5.4.2_3.0_1724035915220.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v2_en_5.4.2_3.0_1724035915220.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_posal_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POSAL_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v4_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v4_en.md new file mode 100644 index 00000000000000..2f7fada8d819b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_posal_v4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_posal_v4 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_posal_v4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_posal_v4` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v4_en_5.4.2_3.0_1724102628437.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_posal_v4_en_5.4.2_3.0_1724102628437.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_posal_v4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_posal_v4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_POSAL_v4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_en.md new file mode 100644 index 00000000000000..5ef92b1bbfcf47 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_en_5.4.2_3.0_1724103290857.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_en_5.4.2_3.0_1724103290857.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_test42_8_RS_SE \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline_en.md new file mode 100644 index 00000000000000..0ca780bc9a5df1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline_en_5.4.2_3.0_1724103462700.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline_en_5.4.2_3.0_1724103462700.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_sapol_test42_8_rs_northern_sami_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SAPOL_test42_8_RS_SE + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_en.md new file mode 100644 index 00000000000000..d5a2dc06e664cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_1024_128 T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_1024_128 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_1024_128` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_en_5.4.2_3.0_1724032824182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_en_5.4.2_3.0_1724032824182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_1024_128","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_1024_128", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_1024_128| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-1024-128 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline_en.md new file mode 100644 index 00000000000000..ebd561bac72c03 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline pipeline T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline_en_5.4.2_3.0_1724032881864.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline_en_5.4.2_3.0_1724032881864.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_1024_128_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-1024-128 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_en.md new file mode 100644 index 00000000000000..b34672a702cb18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_1024_256 T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_1024_256 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_1024_256` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_en_5.4.2_3.0_1724067917349.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_en_5.4.2_3.0_1724067917349.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_1024_256","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_1024_256", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_1024_256| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-1024-256 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline_en.md new file mode 100644 index 00000000000000..fae2f4fedc4ada --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline pipeline T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline_en_5.4.2_3.0_1724067975932.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline_en_5.4.2_3.0_1724067975932.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_1024_256_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-1024-256 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_en.md new file mode 100644 index 00000000000000..677faf7a6cdea5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_en_5.4.2_3.0_1724080411409.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_en_5.4.2_3.0_1724080411409.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-512-128-finetuned-rudaw-ku-512-128-20epochs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline_en.md new file mode 100644 index 00000000000000..b9e8368db2dde7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline pipeline T5Transformer from pedramyamini +author: John Snow Labs +name: kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline` is a English model originally trained by pedramyamini. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline_en_5.4.2_3.0_1724080468596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline_en_5.4.2_3.0_1724080468596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kurdish_t5_base_finetuned_rudaw_kurdish_512_128_finetuned_rudaw_kurdish_512_128_20epochs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/pedramyamini/ku_t5_base-finetuned-rudaw-ku-512-128-finetuned-rudaw-ku-512-128-20epochs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_en.md b/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_en.md new file mode 100644 index 00000000000000..46fb5b28d3c1b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lamini_flan_t5_783m_evolinstruct T5Transformer from KnutJaegersberg +author: John Snow Labs +name: lamini_flan_t5_783m_evolinstruct +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lamini_flan_t5_783m_evolinstruct` is a English model originally trained by KnutJaegersberg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_783m_evolinstruct_en_5.4.2_3.0_1724056264704.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_783m_evolinstruct_en_5.4.2_3.0_1724056264704.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lamini_flan_t5_783m_evolinstruct","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lamini_flan_t5_783m_evolinstruct", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lamini_flan_t5_783m_evolinstruct| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/KnutJaegersberg/LaMini-Flan-T5-783M-EvolInstruct \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_pipeline_en.md new file mode 100644 index 00000000000000..3a870619ecbb89 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lamini_flan_t5_783m_evolinstruct_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lamini_flan_t5_783m_evolinstruct_pipeline pipeline T5Transformer from KnutJaegersberg +author: John Snow Labs +name: lamini_flan_t5_783m_evolinstruct_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lamini_flan_t5_783m_evolinstruct_pipeline` is a English model originally trained by KnutJaegersberg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_783m_evolinstruct_pipeline_en_5.4.2_3.0_1724056396166.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lamini_flan_t5_783m_evolinstruct_pipeline_en_5.4.2_3.0_1724056396166.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lamini_flan_t5_783m_evolinstruct_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lamini_flan_t5_783m_evolinstruct_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lamini_flan_t5_783m_evolinstruct_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/KnutJaegersberg/LaMini-Flan-T5-783M-EvolInstruct + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_en.md new file mode 100644 index 00000000000000..dd47be1a2b031b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_cls_swedish T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_cls_swedish +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_cls_swedish` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_swedish_en_5.4.2_3.0_1724048577392.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_swedish_en_5.4.2_3.0_1724048577392.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_cls_swedish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_cls_swedish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_cls_swedish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|178.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_cls_sv \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_pipeline_en.md new file mode 100644 index 00000000000000..82571b6de331de --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_cls_swedish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_cls_swedish_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_cls_swedish_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_cls_swedish_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_swedish_pipeline_en_5.4.2_3.0_1724048638113.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_cls_swedish_pipeline_en_5.4.2_3.0_1724048638113.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_cls_swedish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_cls_swedish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_cls_swedish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|178.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_cls_sv + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_en.md new file mode 100644 index 00000000000000..0460122d56bbe0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_multitask_german_english T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_german_english +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_german_english` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_german_english_en_5.4.2_3.0_1724095938389.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_german_english_en_5.4.2_3.0_1724095938389.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_multitask_german_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_multitask_german_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_german_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_de_en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_pipeline_en.md new file mode 100644 index 00000000000000..92e4170df6e249 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_german_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_multitask_german_english_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_german_english_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_german_english_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_german_english_pipeline_en_5.4.2_3.0_1724095999304.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_german_english_pipeline_en_5.4.2_3.0_1724095999304.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_multitask_german_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_multitask_german_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_german_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_de_en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_en.md new file mode 100644 index 00000000000000..29c6913b849aa7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_multitask_spanish_italian T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_spanish_italian +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_spanish_italian` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_italian_en_5.4.2_3.0_1724067872247.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_italian_en_5.4.2_3.0_1724067872247.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_multitask_spanish_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_multitask_spanish_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_spanish_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_es_it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_pipeline_en.md new file mode 100644 index 00000000000000..8218c0477c94f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_multitask_spanish_italian_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_spanish_italian_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_spanish_italian_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_italian_pipeline_en_5.4.2_3.0_1724067935989.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_italian_pipeline_en_5.4.2_3.0_1724067935989.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_multitask_spanish_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_multitask_spanish_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_spanish_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_es_it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_en.md new file mode 100644 index 00000000000000..332701f9802442 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_multitask_spanish_swedish T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_spanish_swedish +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_spanish_swedish` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_swedish_en_5.4.2_3.0_1724069306854.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_swedish_en_5.4.2_3.0_1724069306854.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_multitask_spanish_swedish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_multitask_spanish_swedish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_spanish_swedish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_es_sv \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_pipeline_en.md new file mode 100644 index 00000000000000..c9c65dfb12d345 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_multitask_spanish_swedish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_multitask_spanish_swedish_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_multitask_spanish_swedish_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_multitask_spanish_swedish_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_swedish_pipeline_en_5.4.2_3.0_1724069369093.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_multitask_spanish_swedish_pipeline_en_5.4.2_3.0_1724069369093.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_multitask_spanish_swedish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_multitask_spanish_swedish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_multitask_spanish_swedish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_multitask_es_sv + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_en.md new file mode 100644 index 00000000000000..3ff36e18016c31 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_english_italian_small_finetuned T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_english_italian_small_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_english_italian_small_finetuned` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_english_italian_small_finetuned_en_5.4.2_3.0_1724072354643.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_english_italian_small_finetuned_en_5.4.2_3.0_1724072354643.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_english_italian_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_english_italian_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_english_italian_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_en_it_small_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..817c501f4ff8f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_english_italian_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_english_italian_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_english_italian_small_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_english_italian_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_english_italian_small_finetuned_pipeline_en_5.4.2_3.0_1724072416159.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_english_italian_small_finetuned_pipeline_en_5.4.2_3.0_1724072416159.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_english_italian_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_english_italian_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_english_italian_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.6 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_en_it_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_en.md new file mode 100644 index 00000000000000..a68c8816a55574 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_french_swedish_small_finetuned T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_french_swedish_small_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_french_swedish_small_finetuned` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_french_swedish_small_finetuned_en_5.4.2_3.0_1724092268292.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_french_swedish_small_finetuned_en_5.4.2_3.0_1724092268292.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_french_swedish_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_french_swedish_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_french_swedish_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_fr_sv_small_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..5870723fc7b966 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_french_swedish_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_french_swedish_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_french_swedish_small_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_french_swedish_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_french_swedish_small_finetuned_pipeline_en_5.4.2_3.0_1724092328289.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_french_swedish_small_finetuned_pipeline_en_5.4.2_3.0_1724092328289.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_french_swedish_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_french_swedish_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_french_swedish_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_fr_sv_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_en.md new file mode 100644 index 00000000000000..9781d55a1fbd88 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_german_czech_small_finetuned T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_german_czech_small_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_german_czech_small_finetuned` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_german_czech_small_finetuned_en_5.4.2_3.0_1724094859607.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_german_czech_small_finetuned_en_5.4.2_3.0_1724094859607.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_german_czech_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_german_czech_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_german_czech_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.4 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_de_cs_small_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..a880b8a02d78d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_german_czech_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_german_czech_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_german_czech_small_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_german_czech_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_german_czech_small_finetuned_pipeline_en_5.4.2_3.0_1724094919355.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_german_czech_small_finetuned_pipeline_en_5.4.2_3.0_1724094919355.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_german_czech_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_german_czech_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_german_czech_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.4 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_de_cs_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_en.md new file mode 100644 index 00000000000000..a085a3b90fd3d4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_german T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_german +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_german` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_en_5.4.2_3.0_1724074893732.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_en_5.4.2_3.0_1724074893732.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_pipeline_en.md new file mode 100644 index 00000000000000..dfb4a684752a4b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_german_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_german_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_german_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_pipeline_en_5.4.2_3.0_1724074954329.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_pipeline_en_5.4.2_3.0_1724074954329.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_italian_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_italian_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_en.md new file mode 100644 index 00000000000000..9c5f6529be2f7b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_german_small_finetuned T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_german_small_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_german_small_finetuned` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_small_finetuned_en_5.4.2_3.0_1724032512128.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_small_finetuned_en_5.4.2_3.0_1724032512128.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_german_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_italian_german_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_german_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_de_small_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..a493fc3fdbfaf2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_italian_german_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_italian_german_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_italian_german_small_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_italian_german_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_small_finetuned_pipeline_en_5.4.2_3.0_1724032572041.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_italian_german_small_finetuned_pipeline_en_5.4.2_3.0_1724032572041.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_italian_german_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_italian_german_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_italian_german_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_it_de_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_en.md new file mode 100644 index 00000000000000..d8bc851aebb89d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_spanish_italian T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_spanish_italian +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_spanish_italian` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_en_5.4.2_3.0_1724082681675.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_en_5.4.2_3.0_1724082681675.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_spanish_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_spanish_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_spanish_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_es_it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_pipeline_en.md new file mode 100644 index 00000000000000..6db3146c764220 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_spanish_italian_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_spanish_italian_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_spanish_italian_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_pipeline_en_5.4.2_3.0_1724082741953.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_pipeline_en_5.4.2_3.0_1724082741953.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_spanish_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_spanish_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_spanish_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|180.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_es_it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_en.md new file mode 100644 index 00000000000000..e41b4a0a6974f4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_t5_small_trans_spanish_italian_small_finetuned T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_spanish_italian_small_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_spanish_italian_small_finetuned` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_small_finetuned_en_5.4.2_3.0_1724095472159.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_small_finetuned_en_5.4.2_3.0_1724095472159.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_t5_small_trans_spanish_italian_small_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_t5_small_trans_spanish_italian_small_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_spanish_italian_small_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_es_it_small_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..c152a6ae9a4d83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_t5_small_trans_spanish_italian_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_spanish_italian_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_spanish_italian_small_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_spanish_italian_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_small_finetuned_pipeline_en_5.4.2_3.0_1724095532306.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_spanish_italian_small_finetuned_pipeline_en_5.4.2_3.0_1724095532306.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_spanish_italian_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_spanish_italian_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_spanish_italian_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.5 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_es_it_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_en.md new file mode 100644 index 00000000000000..8ae0ed473e51c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English legal_text_simplifier T5Transformer from marcus2000 +author: John Snow Labs +name: legal_text_simplifier +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_text_simplifier` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_text_simplifier_en_5.4.2_3.0_1724045216165.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_text_simplifier_en_5.4.2_3.0_1724045216165.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("legal_text_simplifier","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("legal_text_simplifier", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_text_simplifier| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/marcus2000/legal_text_simplifier \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_pipeline_en.md new file mode 100644 index 00000000000000..9bf38ff598e10d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-legal_text_simplifier_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_text_simplifier_pipeline pipeline T5Transformer from marcus2000 +author: John Snow Labs +name: legal_text_simplifier_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_text_simplifier_pipeline` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_text_simplifier_pipeline_en_5.4.2_3.0_1724045379775.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_text_simplifier_pipeline_en_5.4.2_3.0_1724045379775.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_text_simplifier_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_text_simplifier_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_text_simplifier_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/marcus2000/legal_text_simplifier + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_en.md b/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_en.md new file mode 100644 index 00000000000000..1e957b674fe610 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lesson_summarization_machinelearningzuu T5Transformer from machinelearningzuu +author: John Snow Labs +name: lesson_summarization_machinelearningzuu +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lesson_summarization_machinelearningzuu` is a English model originally trained by machinelearningzuu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lesson_summarization_machinelearningzuu_en_5.4.2_3.0_1724073859117.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lesson_summarization_machinelearningzuu_en_5.4.2_3.0_1724073859117.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lesson_summarization_machinelearningzuu","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lesson_summarization_machinelearningzuu", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lesson_summarization_machinelearningzuu| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|330.8 MB| + +## References + +https://huggingface.co/machinelearningzuu/lesson-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_pipeline_en.md new file mode 100644 index 00000000000000..8813480bd6d2fe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lesson_summarization_machinelearningzuu_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lesson_summarization_machinelearningzuu_pipeline pipeline T5Transformer from machinelearningzuu +author: John Snow Labs +name: lesson_summarization_machinelearningzuu_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lesson_summarization_machinelearningzuu_pipeline` is a English model originally trained by machinelearningzuu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lesson_summarization_machinelearningzuu_pipeline_en_5.4.2_3.0_1724073877266.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lesson_summarization_machinelearningzuu_pipeline_en_5.4.2_3.0_1724073877266.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lesson_summarization_machinelearningzuu_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lesson_summarization_machinelearningzuu_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lesson_summarization_machinelearningzuu_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|330.8 MB| + +## References + +https://huggingface.co/machinelearningzuu/lesson-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_en.md b/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_en.md new file mode 100644 index 00000000000000..8d29f2e7c45e81 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lewit_informal T5Transformer from s-nlp +author: John Snow Labs +name: lewit_informal +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lewit_informal` is a English model originally trained by s-nlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lewit_informal_en_5.4.2_3.0_1724029685269.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lewit_informal_en_5.4.2_3.0_1724029685269.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lewit_informal","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lewit_informal", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lewit_informal| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.5 MB| + +## References + +https://huggingface.co/s-nlp/lewit-informal \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_pipeline_en.md new file mode 100644 index 00000000000000..6416baf089b8d8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lewit_informal_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lewit_informal_pipeline pipeline T5Transformer from s-nlp +author: John Snow Labs +name: lewit_informal_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lewit_informal_pipeline` is a English model originally trained by s-nlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lewit_informal_pipeline_en_5.4.2_3.0_1724029735742.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lewit_informal_pipeline_en_5.4.2_3.0_1724029735742.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lewit_informal_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lewit_informal_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lewit_informal_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.5 MB| + +## References + +https://huggingface.co/s-nlp/lewit-informal + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_en.md new file mode 100644 index 00000000000000..d457229f8b57d3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lil_sumsum T5Transformer from SamuelMiller +author: John Snow Labs +name: lil_sumsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lil_sumsum` is a English model originally trained by SamuelMiller. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lil_sumsum_en_5.4.2_3.0_1724073588072.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lil_sumsum_en_5.4.2_3.0_1724073588072.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lil_sumsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lil_sumsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lil_sumsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/SamuelMiller/lil_sumsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_pipeline_en.md new file mode 100644 index 00000000000000..d77d8780981c29 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-lil_sumsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lil_sumsum_pipeline pipeline T5Transformer from SamuelMiller +author: John Snow Labs +name: lil_sumsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lil_sumsum_pipeline` is a English model originally trained by SamuelMiller. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lil_sumsum_pipeline_en_5.4.2_3.0_1724073605066.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lil_sumsum_pipeline_en_5.4.2_3.0_1724073605066.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lil_sumsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lil_sumsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lil_sumsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/SamuelMiller/lil_sumsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_en.md new file mode 100644 index 00000000000000..2be15bdc05fdc5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English live_t5_base T5Transformer from RUCAIBox +author: John Snow Labs +name: live_t5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`live_t5_base` is a English model originally trained by RUCAIBox. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/live_t5_base_en_5.4.2_3.0_1724080701353.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/live_t5_base_en_5.4.2_3.0_1724080701353.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("live_t5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("live_t5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|live_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|624.3 MB| + +## References + +https://huggingface.co/RUCAIBox/live-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_pipeline_en.md new file mode 100644 index 00000000000000..eb552134429c22 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-live_t5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English live_t5_base_pipeline pipeline T5Transformer from RUCAIBox +author: John Snow Labs +name: live_t5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`live_t5_base_pipeline` is a English model originally trained by RUCAIBox. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/live_t5_base_pipeline_en_5.4.2_3.0_1724080848141.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/live_t5_base_pipeline_en_5.4.2_3.0_1724080848141.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("live_t5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("live_t5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|live_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|624.3 MB| + +## References + +https://huggingface.co/RUCAIBox/live-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_en.md new file mode 100644 index 00000000000000..afbc0fc5b970bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum T5Transformer from rooftopcoder +author: John Snow Labs +name: long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_en_5.4.2_3.0_1724061281741.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_en_5.4.2_3.0_1724061281741.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 MB| + +## References + +https://huggingface.co/rooftopcoder/long-t5-tglobal-base-16384-book-summary-finetuned-dialogsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline_en.md new file mode 100644 index 00000000000000..d4519d52e6e5d4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline pipeline T5Transformer from rooftopcoder +author: John Snow Labs +name: long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline_en_5.4.2_3.0_1724061284405.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline_en_5.4.2_3.0_1724061284405.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|long_t5_tglobal_base_16384_book_summary_finetuned_dialogsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.5 MB| + +## References + +https://huggingface.co/rooftopcoder/long-t5-tglobal-base-16384-book-summary-finetuned-dialogsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_en.md b/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_en.md new file mode 100644 index 00000000000000..a4694b4416c01d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English machine_translation_quynhit T5Transformer from quynhit +author: John Snow Labs +name: machine_translation_quynhit +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`machine_translation_quynhit` is a English model originally trained by quynhit. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/machine_translation_quynhit_en_5.4.2_3.0_1724070080537.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/machine_translation_quynhit_en_5.4.2_3.0_1724070080537.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("machine_translation_quynhit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("machine_translation_quynhit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|machine_translation_quynhit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.4 MB| + +## References + +https://huggingface.co/quynhit/machine-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_pipeline_en.md new file mode 100644 index 00000000000000..1236143bf9dc44 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-machine_translation_quynhit_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English machine_translation_quynhit_pipeline pipeline T5Transformer from quynhit +author: John Snow Labs +name: machine_translation_quynhit_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`machine_translation_quynhit_pipeline` is a English model originally trained by quynhit. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/machine_translation_quynhit_pipeline_en_5.4.2_3.0_1724070132140.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/machine_translation_quynhit_pipeline_en_5.4.2_3.0_1724070132140.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("machine_translation_quynhit_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("machine_translation_quynhit_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|machine_translation_quynhit_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.4 MB| + +## References + +https://huggingface.co/quynhit/machine-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_en.md b/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_en.md new file mode 100644 index 00000000000000..a1494baaef5359 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112 T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_en_5.4.2_3.0_1724065826228.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_en_5.4.2_3.0_1724065826228.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline_en.md new file mode 100644 index 00000000000000..2d88e006bf62bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline_en_5.4.2_3.0_1724065977150.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline_en_5.4.2_3.0_1724065977150.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_base_boun_split_second_v1_retrain_on_second_imst_v2_1112 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mestrado2_en.md b/docs/_posts/ahmedlone127/2024-08-19-mestrado2_en.md new file mode 100644 index 00000000000000..42dbc71fab8015 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mestrado2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mestrado2 T5Transformer from clertonalmeida +author: John Snow Labs +name: mestrado2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mestrado2` is a English model originally trained by clertonalmeida. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mestrado2_en_5.4.2_3.0_1724045218237.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mestrado2_en_5.4.2_3.0_1724045218237.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mestrado2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mestrado2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mestrado2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|313.0 MB| + +## References + +https://huggingface.co/clertonalmeida/mestrado2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mestrado2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mestrado2_pipeline_en.md new file mode 100644 index 00000000000000..e384f9ac11c875 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mestrado2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mestrado2_pipeline pipeline T5Transformer from clertonalmeida +author: John Snow Labs +name: mestrado2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mestrado2_pipeline` is a English model originally trained by clertonalmeida. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mestrado2_pipeline_en_5.4.2_3.0_1724045241271.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mestrado2_pipeline_en_5.4.2_3.0_1724045241271.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mestrado2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mestrado2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mestrado2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|313.0 MB| + +## References + +https://huggingface.co/clertonalmeida/mestrado2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model300t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-model300t5_en.md new file mode 100644 index 00000000000000..3f4dc8309fee0d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model300t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model300t5 T5Transformer from Saddammm +author: John Snow Labs +name: model300t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model300t5` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model300t5_en_5.4.2_3.0_1724109891759.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model300t5_en_5.4.2_3.0_1724109891759.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model300t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model300t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model300t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/Model300T5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model300t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-model300t5_pipeline_en.md new file mode 100644 index 00000000000000..8653f4fb44e090 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model300t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model300t5_pipeline pipeline T5Transformer from Saddammm +author: John Snow Labs +name: model300t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model300t5_pipeline` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model300t5_pipeline_en_5.4.2_3.0_1724109940928.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model300t5_pipeline_en_5.4.2_3.0_1724109940928.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model300t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model300t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model300t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/Model300T5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_en.md new file mode 100644 index 00000000000000..aa7531ec42dcff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model_eldestructo T5Transformer from ElDestructo +author: John Snow Labs +name: model_eldestructo +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_eldestructo` is a English model originally trained by ElDestructo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_eldestructo_en_5.4.2_3.0_1724070013169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_eldestructo_en_5.4.2_3.0_1724070013169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model_eldestructo","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model_eldestructo", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_eldestructo| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/ElDestructo/model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_pipeline_en.md new file mode 100644 index 00000000000000..8a3cfbe5336ab6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_eldestructo_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model_eldestructo_pipeline pipeline T5Transformer from ElDestructo +author: John Snow Labs +name: model_eldestructo_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_eldestructo_pipeline` is a English model originally trained by ElDestructo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_eldestructo_pipeline_en_5.4.2_3.0_1724070158006.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_eldestructo_pipeline_en_5.4.2_3.0_1724070158006.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model_eldestructo_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model_eldestructo_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_eldestructo_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/ElDestructo/model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_en.md new file mode 100644 index 00000000000000..efd4cb64713836 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model_pre_tuning T5Transformer from Otavares +author: John Snow Labs +name: model_pre_tuning +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_pre_tuning` is a English model originally trained by Otavares. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_pre_tuning_en_5.4.2_3.0_1724108205892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_pre_tuning_en_5.4.2_3.0_1724108205892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model_pre_tuning","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model_pre_tuning", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_pre_tuning| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Otavares/model_pre_tuning \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_pipeline_en.md new file mode 100644 index 00000000000000..160cd0c784d9e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_pre_tuning_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model_pre_tuning_pipeline pipeline T5Transformer from Otavares +author: John Snow Labs +name: model_pre_tuning_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_pre_tuning_pipeline` is a English model originally trained by Otavares. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_pre_tuning_pipeline_en_5.4.2_3.0_1724108253552.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_pre_tuning_pipeline_en_5.4.2_3.0_1724108253552.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model_pre_tuning_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model_pre_tuning_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_pre_tuning_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Otavares/model_pre_tuning + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_soroush_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_soroush_en.md new file mode 100644 index 00000000000000..2eecbb5d799cc1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_soroush_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English model_soroush T5Transformer from soroush +author: John Snow Labs +name: model_soroush +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_soroush` is a English model originally trained by soroush. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_soroush_en_5.4.2_3.0_1724070619541.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_soroush_en_5.4.2_3.0_1724070619541.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("model_soroush","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("model_soroush", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_soroush| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/soroush/model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-model_soroush_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-model_soroush_pipeline_en.md new file mode 100644 index 00000000000000..41cb08d69548e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-model_soroush_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English model_soroush_pipeline pipeline T5Transformer from soroush +author: John Snow Labs +name: model_soroush_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`model_soroush_pipeline` is a English model originally trained by soroush. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/model_soroush_pipeline_en_5.4.2_3.0_1724070669159.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/model_soroush_pipeline_en_5.4.2_3.0_1724070669159.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("model_soroush_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("model_soroush_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|model_soroush_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/soroush/model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_en.md b/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_en.md new file mode 100644 index 00000000000000..4325e7ee61df3a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English molt5_small_fine_tune T5Transformer from Phudish +author: John Snow Labs +name: molt5_small_fine_tune +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`molt5_small_fine_tune` is a English model originally trained by Phudish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/molt5_small_fine_tune_en_5.4.2_3.0_1724078220648.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/molt5_small_fine_tune_en_5.4.2_3.0_1724078220648.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("molt5_small_fine_tune","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("molt5_small_fine_tune", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|molt5_small_fine_tune| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/Phudish/molt5-small-fine-tune \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_pipeline_en.md new file mode 100644 index 00000000000000..964d977a42a594 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-molt5_small_fine_tune_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English molt5_small_fine_tune_pipeline pipeline T5Transformer from Phudish +author: John Snow Labs +name: molt5_small_fine_tune_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`molt5_small_fine_tune_pipeline` is a English model originally trained by Phudish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/molt5_small_fine_tune_pipeline_en_5.4.2_3.0_1724078237318.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/molt5_small_fine_tune_pipeline_en_5.4.2_3.0_1724078237318.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("molt5_small_fine_tune_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("molt5_small_fine_tune_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|molt5_small_fine_tune_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/Phudish/molt5-small-fine-tune + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_en.md new file mode 100644 index 00000000000000..c9370d1bd5feb3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mrpc_t5_base_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: mrpc_t5_base_seed_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mrpc_t5_base_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mrpc_t5_base_seed_1_en_5.4.2_3.0_1724092558232.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mrpc_t5_base_seed_1_en_5.4.2_3.0_1724092558232.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mrpc_t5_base_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mrpc_t5_base_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mrpc_t5_base_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|926.0 MB| + +## References + +https://huggingface.co/utahnlp/mrpc_t5-base_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..832aa04ef80615 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mrpc_t5_base_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mrpc_t5_base_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: mrpc_t5_base_seed_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mrpc_t5_base_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mrpc_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724092629227.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mrpc_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724092629227.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mrpc_t5_base_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mrpc_t5_base_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mrpc_t5_base_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|926.0 MB| + +## References + +https://huggingface.co/utahnlp/mrpc_t5-base_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_en.md new file mode 100644 index 00000000000000..255bf475c8b2a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_600m_flores200_baseline T5Transformer from hlillemark +author: John Snow Labs +name: mt5_600m_flores200_baseline +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_600m_flores200_baseline` is a English model originally trained by hlillemark. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_600m_flores200_baseline_en_5.4.2_3.0_1724056319729.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_600m_flores200_baseline_en_5.4.2_3.0_1724056319729.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_600m_flores200_baseline","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_600m_flores200_baseline", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_600m_flores200_baseline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/hlillemark/mt5-600M-flores200-baseline \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_pipeline_en.md new file mode 100644 index 00000000000000..067b9ce0fdf2b7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_600m_flores200_baseline_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_600m_flores200_baseline_pipeline pipeline T5Transformer from hlillemark +author: John Snow Labs +name: mt5_600m_flores200_baseline_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_600m_flores200_baseline_pipeline` is a English model originally trained by hlillemark. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_600m_flores200_baseline_pipeline_en_5.4.2_3.0_1724056799008.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_600m_flores200_baseline_pipeline_en_5.4.2_3.0_1724056799008.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_600m_flores200_baseline_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_600m_flores200_baseline_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_600m_flores200_baseline_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/hlillemark/mt5-600M-flores200-baseline + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_en.md new file mode 100644 index 00000000000000..5b7044a7ea02c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_100000 T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_100000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_100000` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_100000_en_5.4.2_3.0_1724094094949.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_100000_en_5.4.2_3.0_1724094094949.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_100000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_100000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_100000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_100000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_pipeline_en.md new file mode 100644 index 00000000000000..ba4377ceded141 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_100000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_100000_pipeline pipeline T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_100000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_100000_pipeline` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_100000_pipeline_en_5.4.2_3.0_1724094292756.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_100000_pipeline_en_5.4.2_3.0_1724094292756.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_100000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_100000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_100000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_100000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_en.md new file mode 100644 index 00000000000000..797ee212a7dcaf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_all_100000 T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_all_100000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_all_100000` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_all_100000_en_5.4.2_3.0_1724037278031.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_all_100000_en_5.4.2_3.0_1724037278031.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_all_100000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_all_100000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_all_100000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_all_100000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_pipeline_en.md new file mode 100644 index 00000000000000..7f5061c0932da0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_all_100000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_all_100000_pipeline pipeline T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_base_all_100000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_all_100000_pipeline` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_all_100000_pipeline_en_5.4.2_3.0_1724037463436.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_all_100000_pipeline_en_5.4.2_3.0_1724037463436.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_all_100000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_all_100000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_all_100000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.6 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_base_all_100000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_cs.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_cs.md new file mode 100644 index 00000000000000..d678477b2b775d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_cs.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Czech mt5_base_binary_czech_iiia_02c T5Transformer from chi2024 +author: John Snow Labs +name: mt5_base_binary_czech_iiia_02c +date: 2024-08-19 +tags: [cs, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: cs +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_binary_czech_iiia_02c` is a Czech model originally trained by chi2024. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_binary_czech_iiia_02c_cs_5.4.2_3.0_1724035520734.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_binary_czech_iiia_02c_cs_5.4.2_3.0_1724035520734.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_binary_czech_iiia_02c","cs") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_binary_czech_iiia_02c", "cs") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_binary_czech_iiia_02c| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|cs| +|Size:|2.3 GB| + +## References + +https://huggingface.co/chi2024/mt5-base-binary-cs-iiia-02c \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_pipeline_cs.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_pipeline_cs.md new file mode 100644 index 00000000000000..9ea02ee6e71c2c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_binary_czech_iiia_02c_pipeline_cs.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Czech mt5_base_binary_czech_iiia_02c_pipeline pipeline T5Transformer from chi2024 +author: John Snow Labs +name: mt5_base_binary_czech_iiia_02c_pipeline +date: 2024-08-19 +tags: [cs, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: cs +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_binary_czech_iiia_02c_pipeline` is a Czech model originally trained by chi2024. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_binary_czech_iiia_02c_pipeline_cs_5.4.2_3.0_1724035810176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_binary_czech_iiia_02c_pipeline_cs_5.4.2_3.0_1724035810176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_binary_czech_iiia_02c_pipeline", lang = "cs") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_binary_czech_iiia_02c_pipeline", lang = "cs") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_binary_czech_iiia_02c_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|cs| +|Size:|2.3 GB| + +## References + +https://huggingface.co/chi2024/mt5-base-binary-cs-iiia-02c + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_esquad_ae_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_esquad_ae_pipeline_es.md new file mode 100644 index 00000000000000..dad228092a5c5c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_esquad_ae_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_base_esquad_ae_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_esquad_ae_pipeline +date: 2024-08-19 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_esquad_ae_pipeline` is a Castilian, Spanish model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_esquad_ae_pipeline_es_5.4.2_3.0_1724038997598.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_esquad_ae_pipeline_es_5.4.2_3.0_1724038997598.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_esquad_ae_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_esquad_ae_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_esquad_ae_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|2.4 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-esquad-ae + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister_en.md new file mode 100644 index 00000000000000..599c71a29234c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister T5Transformer from himanshubeniwal +author: John Snow Labs +name: mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister` is a English model originally trained by himanshubeniwal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister_en_5.4.2_3.0_1724044360185.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister_en_5.4.2_3.0_1724044360185.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_kazakh_tonga_tonga_islands_english_filthy_primeminister| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/himanshubeniwal/mt5-base-finetuned-kk-to-en-filthy-PrimeMinister \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_ar.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_ar.md new file mode 100644 index 00000000000000..62f127c57ee64f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_ar.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Arabic mt5_base_finetuned_summarization_arabic T5Transformer from shahadalll +author: John Snow Labs +name: mt5_base_finetuned_summarization_arabic +date: 2024-08-19 +tags: [ar, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ar +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_summarization_arabic` is a Arabic model originally trained by shahadalll. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_summarization_arabic_ar_5.4.2_3.0_1724085881566.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_summarization_arabic_ar_5.4.2_3.0_1724085881566.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_summarization_arabic","ar") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_summarization_arabic", "ar") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_summarization_arabic| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ar| +|Size:|2.3 GB| + +## References + +https://huggingface.co/shahadalll/mt5-base-finetuned-summarization-Arabic \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_pipeline_ar.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_pipeline_ar.md new file mode 100644 index 00000000000000..d6d9972c4503b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_summarization_arabic_pipeline_ar.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Arabic mt5_base_finetuned_summarization_arabic_pipeline pipeline T5Transformer from shahadalll +author: John Snow Labs +name: mt5_base_finetuned_summarization_arabic_pipeline +date: 2024-08-19 +tags: [ar, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ar +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_summarization_arabic_pipeline` is a Arabic model originally trained by shahadalll. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_summarization_arabic_pipeline_ar_5.4.2_3.0_1724086114752.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_summarization_arabic_pipeline_ar_5.4.2_3.0_1724086114752.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_finetuned_summarization_arabic_pipeline", lang = "ar") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_finetuned_summarization_arabic_pipeline", lang = "ar") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_summarization_arabic_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ar| +|Size:|2.3 GB| + +## References + +https://huggingface.co/shahadalll/mt5-base-finetuned-summarization-Arabic + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_en.md new file mode 100644 index 00000000000000..f09bf3b267c856 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_test_30483_prefix_summarize T5Transformer from emilstabil +author: John Snow Labs +name: mt5_base_finetuned_test_30483_prefix_summarize +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_test_30483_prefix_summarize` is a English model originally trained by emilstabil. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_test_30483_prefix_summarize_en_5.4.2_3.0_1724080133329.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_test_30483_prefix_summarize_en_5.4.2_3.0_1724080133329.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_test_30483_prefix_summarize","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_test_30483_prefix_summarize", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_test_30483_prefix_summarize| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/emilstabil/mt5-base-finetuned-test_30483_prefix_summarize \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_pipeline_en.md new file mode 100644 index 00000000000000..d04086675fdcbf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_test_30483_prefix_summarize_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_finetuned_test_30483_prefix_summarize_pipeline pipeline T5Transformer from emilstabil +author: John Snow Labs +name: mt5_base_finetuned_test_30483_prefix_summarize_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_test_30483_prefix_summarize_pipeline` is a English model originally trained by emilstabil. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_test_30483_prefix_summarize_pipeline_en_5.4.2_3.0_1724080309417.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_test_30483_prefix_summarize_pipeline_en_5.4.2_3.0_1724080309417.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_finetuned_test_30483_prefix_summarize_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_finetuned_test_30483_prefix_summarize_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_test_30483_prefix_summarize_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/emilstabil/mt5-base-finetuned-test_30483_prefix_summarize + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_pipeline_uk.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_pipeline_uk.md new file mode 100644 index 00000000000000..ca7a3862ebfa9c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_pipeline_uk.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Ukrainian mt5_base_finetuned_ukr_gec_pipeline pipeline T5Transformer from schhwmn +author: John Snow Labs +name: mt5_base_finetuned_ukr_gec_pipeline +date: 2024-08-19 +tags: [uk, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_ukr_gec_pipeline` is a Ukrainian model originally trained by schhwmn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_ukr_gec_pipeline_uk_5.4.2_3.0_1724034238685.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_ukr_gec_pipeline_uk_5.4.2_3.0_1724034238685.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_finetuned_ukr_gec_pipeline", lang = "uk") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_finetuned_ukr_gec_pipeline", lang = "uk") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_ukr_gec_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|uk| +|Size:|2.3 GB| + +## References + +https://huggingface.co/schhwmn/mt5-base-finetuned-ukr-gec + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_uk.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_uk.md new file mode 100644 index 00000000000000..d67848f4894781 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_finetuned_ukr_gec_uk.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Ukrainian mt5_base_finetuned_ukr_gec T5Transformer from schhwmn +author: John Snow Labs +name: mt5_base_finetuned_ukr_gec +date: 2024-08-19 +tags: [uk, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: uk +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_ukr_gec` is a Ukrainian model originally trained by schhwmn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_ukr_gec_uk_5.4.2_3.0_1724034092802.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_ukr_gec_uk_5.4.2_3.0_1724034092802.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_ukr_gec","uk") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_ukr_gec", "uk") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_ukr_gec| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|uk| +|Size:|2.3 GB| + +## References + +https://huggingface.co/schhwmn/mt5-base-finetuned-ukr-gec \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_en.md new file mode 100644 index 00000000000000..9f4e1763063c3d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_ft_spider_hebrew T5Transformer from yonataneshel +author: John Snow Labs +name: mt5_base_ft_spider_hebrew +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ft_spider_hebrew` is a English model originally trained by yonataneshel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ft_spider_hebrew_en_5.4.2_3.0_1724068666882.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ft_spider_hebrew_en_5.4.2_3.0_1724068666882.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_ft_spider_hebrew","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_ft_spider_hebrew", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ft_spider_hebrew| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/yonataneshel/mt5_base_ft_spider_hebrew \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_pipeline_en.md new file mode 100644 index 00000000000000..79403a0017d019 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ft_spider_hebrew_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_ft_spider_hebrew_pipeline pipeline T5Transformer from yonataneshel +author: John Snow Labs +name: mt5_base_ft_spider_hebrew_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ft_spider_hebrew_pipeline` is a English model originally trained by yonataneshel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ft_spider_hebrew_pipeline_en_5.4.2_3.0_1724068969969.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ft_spider_hebrew_pipeline_en_5.4.2_3.0_1724068969969.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_ft_spider_hebrew_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_ft_spider_hebrew_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ft_spider_hebrew_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/yonataneshel/mt5_base_ft_spider_hebrew + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_it.md new file mode 100644 index 00000000000000..774941ef410b39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_it.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Italian mt5_base_ilgiornale_tonga_tonga_islands_repubblica T5Transformer from gsarti +author: John Snow Labs +name: mt5_base_ilgiornale_tonga_tonga_islands_repubblica +date: 2024-08-19 +tags: [it, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ilgiornale_tonga_tonga_islands_repubblica` is a Italian model originally trained by gsarti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ilgiornale_tonga_tonga_islands_repubblica_it_5.4.2_3.0_1724056612435.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ilgiornale_tonga_tonga_islands_repubblica_it_5.4.2_3.0_1724056612435.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_ilgiornale_tonga_tonga_islands_repubblica","it") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_ilgiornale_tonga_tonga_islands_repubblica", "it") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ilgiornale_tonga_tonga_islands_repubblica| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|it| +|Size:|2.4 GB| + +## References + +https://huggingface.co/gsarti/mt5-base-ilgiornale-to-repubblica \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline_it.md new file mode 100644 index 00000000000000..6e859580b39e45 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline_it.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Italian mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline pipeline T5Transformer from gsarti +author: John Snow Labs +name: mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline +date: 2024-08-19 +tags: [it, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline` is a Italian model originally trained by gsarti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline_it_5.4.2_3.0_1724056853806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline_it_5.4.2_3.0_1724056853806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline", lang = "it") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline", lang = "it") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_ilgiornale_tonga_tonga_islands_repubblica_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|it| +|Size:|2.4 GB| + +## References + +https://huggingface.co/gsarti/mt5-base-ilgiornale-to-repubblica + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_koquad_ae_pipeline_ko.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_koquad_ae_pipeline_ko.md new file mode 100644 index 00000000000000..7b1ba053ef8e74 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_koquad_ae_pipeline_ko.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Korean mt5_base_koquad_ae_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_koquad_ae_pipeline +date: 2024-08-19 +tags: [ko, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_koquad_ae_pipeline` is a Korean model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_koquad_ae_pipeline_ko_5.4.2_3.0_1724031369128.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_koquad_ae_pipeline_ko_5.4.2_3.0_1724031369128.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_koquad_ae_pipeline", lang = "ko") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_koquad_ae_pipeline", lang = "ko") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_koquad_ae_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ko| +|Size:|2.3 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-koquad-ae + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_en.md new file mode 100644 index 00000000000000..ba0bfdfcfe2203 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_poll_generation T5Transformer from Pedrambbk +author: John Snow Labs +name: mt5_base_poll_generation +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_poll_generation` is a English model originally trained by Pedrambbk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_poll_generation_en_5.4.2_3.0_1724036512268.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_poll_generation_en_5.4.2_3.0_1724036512268.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_poll_generation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_poll_generation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_poll_generation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Pedrambbk/mt5-base-poll-generation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_pipeline_en.md new file mode 100644 index 00000000000000..34b51de5e4f146 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_poll_generation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_poll_generation_pipeline pipeline T5Transformer from Pedrambbk +author: John Snow Labs +name: mt5_base_poll_generation_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_poll_generation_pipeline` is a English model originally trained by Pedrambbk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_poll_generation_pipeline_en_5.4.2_3.0_1724036777084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_poll_generation_pipeline_en_5.4.2_3.0_1724036777084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_poll_generation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_poll_generation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_poll_generation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/Pedrambbk/mt5-base-poll-generation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_en.md new file mode 100644 index 00000000000000..4bb8aa78e26f85 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_qg_aap_oficial T5Transformer from tiagoblima +author: John Snow Labs +name: mt5_base_qg_aap_oficial +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_qg_aap_oficial` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_qg_aap_oficial_en_5.4.2_3.0_1724097813051.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_qg_aap_oficial_en_5.4.2_3.0_1724097813051.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_qg_aap_oficial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_qg_aap_oficial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_qg_aap_oficial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/tiagoblima/mt5_base-qg-aap-oficial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_pipeline_en.md new file mode 100644 index 00000000000000..794b575c9b759f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_aap_oficial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_qg_aap_oficial_pipeline pipeline T5Transformer from tiagoblima +author: John Snow Labs +name: mt5_base_qg_aap_oficial_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_qg_aap_oficial_pipeline` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_qg_aap_oficial_pipeline_en_5.4.2_3.0_1724098091691.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_qg_aap_oficial_pipeline_en_5.4.2_3.0_1724098091691.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_qg_aap_oficial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_qg_aap_oficial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_qg_aap_oficial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/tiagoblima/mt5_base-qg-aap-oficial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_en.md new file mode 100644 index 00000000000000..d94b170a4f7d2a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_qg_ap_oficial T5Transformer from tiagoblima +author: John Snow Labs +name: mt5_base_qg_ap_oficial +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_qg_ap_oficial` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_qg_ap_oficial_en_5.4.2_3.0_1724074060084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_qg_ap_oficial_en_5.4.2_3.0_1724074060084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_qg_ap_oficial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_qg_ap_oficial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_qg_ap_oficial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/tiagoblima/mt5_base-qg-ap-oficial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_pipeline_en.md new file mode 100644 index 00000000000000..fa1673ae6837ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_qg_ap_oficial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_qg_ap_oficial_pipeline pipeline T5Transformer from tiagoblima +author: John Snow Labs +name: mt5_base_qg_ap_oficial_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_qg_ap_oficial_pipeline` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_qg_ap_oficial_pipeline_en_5.4.2_3.0_1724074343527.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_qg_ap_oficial_pipeline_en_5.4.2_3.0_1724074343527.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_qg_ap_oficial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_qg_ap_oficial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_qg_ap_oficial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/tiagoblima/mt5_base-qg-ap-oficial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_it.md new file mode 100644 index 00000000000000..95588b916b0415 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_it.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Italian mt5_base_question_answering T5Transformer from gsarti +author: John Snow Labs +name: mt5_base_question_answering +date: 2024-08-19 +tags: [it, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_question_answering` is a Italian model originally trained by gsarti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_question_answering_it_5.4.2_3.0_1724047831460.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_question_answering_it_5.4.2_3.0_1724047831460.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_question_answering","it") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_question_answering", "it") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_question_answering| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|it| +|Size:|2.4 GB| + +## References + +https://huggingface.co/gsarti/mt5-base-question-answering \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_pipeline_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_pipeline_it.md new file mode 100644 index 00000000000000..e15caeeb63964b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_question_answering_pipeline_it.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Italian mt5_base_question_answering_pipeline pipeline T5Transformer from gsarti +author: John Snow Labs +name: mt5_base_question_answering_pipeline +date: 2024-08-19 +tags: [it, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_question_answering_pipeline` is a Italian model originally trained by gsarti. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_question_answering_pipeline_it_5.4.2_3.0_1724048106014.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_question_answering_pipeline_it_5.4.2_3.0_1724048106014.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_question_answering_pipeline", lang = "it") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_question_answering_pipeline", lang = "it") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_question_answering_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|it| +|Size:|2.4 GB| + +## References + +https://huggingface.co/gsarti/mt5-base-question-answering + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_en.md new file mode 100644 index 00000000000000..098dbc2c66e6d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained T5Transformer from parinzee +author: John Snow Labs +name: mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained` is a English model originally trained by parinzee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_en_5.4.2_3.0_1724066475090.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_en_5.4.2_3.0_1724066475090.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/parinzee/mt5-base-thai-multiple-e2e-qg-aug-numsep-retrained \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline_en.md new file mode 100644 index 00000000000000..04b8e4de2e6ab1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline pipeline T5Transformer from parinzee +author: John Snow Labs +name: mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline` is a English model originally trained by parinzee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline_en_5.4.2_3.0_1724066646151.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline_en_5.4.2_3.0_1724066646151.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_thai_multiple_e2e_qg_aug_numsep_retrained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/parinzee/mt5-base-thai-multiple-e2e-qg-aug-numsep-retrained + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_en.md new file mode 100644 index 00000000000000..c7e8e0f97ee518 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_trimmed_italian_90000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_italian_90000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_italian_90000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_90000_en_5.4.2_3.0_1724040579904.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_90000_en_5.4.2_3.0_1724040579904.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_trimmed_italian_90000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_trimmed_italian_90000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_italian_90000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|777.6 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-it-90000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_pipeline_en.md new file mode 100644 index 00000000000000..87124d35929e2a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_italian_90000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_trimmed_italian_90000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_italian_90000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_italian_90000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_90000_pipeline_en_5.4.2_3.0_1724040836238.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_90000_pipeline_en_5.4.2_3.0_1724040836238.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_trimmed_italian_90000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_trimmed_italian_90000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_italian_90000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|777.6 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-it-90000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_japanese_60000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_japanese_60000_en.md new file mode 100644 index 00000000000000..a5e76d781dd83e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_trimmed_japanese_60000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_trimmed_japanese_60000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_japanese_60000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_japanese_60000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_japanese_60000_en_5.4.2_3.0_1724111891164.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_japanese_60000_en_5.4.2_3.0_1724111891164.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_trimmed_japanese_60000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_trimmed_japanese_60000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_japanese_60000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|645.5 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-ja-60000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_en.md new file mode 100644 index 00000000000000..51f50e4f86bff1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_turkish T5Transformer from Mursel +author: John Snow Labs +name: mt5_base_turkish +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_turkish` is a English model originally trained by Mursel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_turkish_en_5.4.2_3.0_1724108104002.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_turkish_en_5.4.2_3.0_1724108104002.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_turkish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_turkish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_turkish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|511.6 MB| + +## References + +https://huggingface.co/Mursel/mt5-base-turkish \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_pipeline_en.md new file mode 100644 index 00000000000000..e48424a28582c1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_base_turkish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_turkish_pipeline pipeline T5Transformer from Mursel +author: John Snow Labs +name: mt5_base_turkish_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_turkish_pipeline` is a English model originally trained by Mursel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_turkish_pipeline_en_5.4.2_3.0_1724108281433.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_turkish_pipeline_en_5.4.2_3.0_1724108281433.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_turkish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_turkish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_turkish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|511.6 MB| + +## References + +https://huggingface.co/Mursel/mt5-base-turkish + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_en.md new file mode 100644 index 00000000000000..18b12196b1527a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_extend_py2500 T5Transformer from Roy029 +author: John Snow Labs +name: mt5_extend_py2500 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_extend_py2500` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_extend_py2500_en_5.4.2_3.0_1724057534678.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_extend_py2500_en_5.4.2_3.0_1724057534678.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_extend_py2500","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_extend_py2500", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_extend_py2500| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_extend_py2500 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_pipeline_en.md new file mode 100644 index 00000000000000..ee0922afc9a613 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_extend_py2500_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_extend_py2500_pipeline pipeline T5Transformer from Roy029 +author: John Snow Labs +name: mt5_extend_py2500_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_extend_py2500_pipeline` is a English model originally trained by Roy029. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_extend_py2500_pipeline_en_5.4.2_3.0_1724057716485.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_extend_py2500_pipeline_en_5.4.2_3.0_1724057716485.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_extend_py2500_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_extend_py2500_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_extend_py2500_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Roy029/mt5_extend_py2500 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_en.md new file mode 100644 index 00000000000000..a91d740661d191 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_hihi T5Transformer from balaramas +author: John Snow Labs +name: mt5_hihi +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_hihi` is a English model originally trained by balaramas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_hihi_en_5.4.2_3.0_1724095243500.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_hihi_en_5.4.2_3.0_1724095243500.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_hihi","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_hihi", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_hihi| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/balaramas/mt5-hihi \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_pipeline_en.md new file mode 100644 index 00000000000000..b76a2680a3430b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_hihi_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_hihi_pipeline pipeline T5Transformer from balaramas +author: John Snow Labs +name: mt5_hihi_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_hihi_pipeline` is a English model originally trained by balaramas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_hihi_pipeline_en_5.4.2_3.0_1724095430666.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_hihi_pipeline_en_5.4.2_3.0_1724095430666.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_hihi_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_hihi_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_hihi_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/balaramas/mt5-hihi + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_no.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_no.md new file mode 100644 index 00000000000000..ea0a2d77888c25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_no.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Norwegian mt5_large_norwegian_info_extraction_3000 T5Transformer from norkart +author: John Snow Labs +name: mt5_large_norwegian_info_extraction_3000 +date: 2024-08-19 +tags: ["no", open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: "no" +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_large_norwegian_info_extraction_3000` is a Norwegian model originally trained by norkart. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_large_norwegian_info_extraction_3000_no_5.4.2_3.0_1724063661919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_large_norwegian_info_extraction_3000_no_5.4.2_3.0_1724063661919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_large_norwegian_info_extraction_3000","no") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_large_norwegian_info_extraction_3000", "no") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_large_norwegian_info_extraction_3000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|no| +|Size:|3.0 GB| + +## References + +https://huggingface.co/norkart/mt5-large-no-info-extraction-3000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_pipeline_no.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_pipeline_no.md new file mode 100644 index 00000000000000..4e8a36e4e651d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_large_norwegian_info_extraction_3000_pipeline_no.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Norwegian mt5_large_norwegian_info_extraction_3000_pipeline pipeline T5Transformer from norkart +author: John Snow Labs +name: mt5_large_norwegian_info_extraction_3000_pipeline +date: 2024-08-19 +tags: ["no", open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: "no" +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_large_norwegian_info_extraction_3000_pipeline` is a Norwegian model originally trained by norkart. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_large_norwegian_info_extraction_3000_pipeline_no_5.4.2_3.0_1724063800531.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_large_norwegian_info_extraction_3000_pipeline_no_5.4.2_3.0_1724063800531.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_large_norwegian_info_extraction_3000_pipeline", lang = "no") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_large_norwegian_info_extraction_3000_pipeline", lang = "no") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_large_norwegian_info_extraction_3000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|no| +|Size:|3.0 GB| + +## References + +https://huggingface.co/norkart/mt5-large-no-info-extraction-3000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_en.md new file mode 100644 index 00000000000000..685288840d46e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_mustc_french T5Transformer from enimai +author: John Snow Labs +name: mt5_mustc_french +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_mustc_french` is a English model originally trained by enimai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_mustc_french_en_5.4.2_3.0_1724065150871.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_mustc_french_en_5.4.2_3.0_1724065150871.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_mustc_french","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_mustc_french", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_mustc_french| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/enimai/mt5-mustc-fr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_pipeline_en.md new file mode 100644 index 00000000000000..d8c36f9c04e820 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_mustc_french_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_mustc_french_pipeline pipeline T5Transformer from enimai +author: John Snow Labs +name: mt5_mustc_french_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_mustc_french_pipeline` is a English model originally trained by enimai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_mustc_french_pipeline_en_5.4.2_3.0_1724065414239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_mustc_french_pipeline_en_5.4.2_3.0_1724065414239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_mustc_french_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_mustc_french_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_mustc_french_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/enimai/mt5-mustc-fr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_pipeline_tr.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_pipeline_tr.md new file mode 100644 index 00000000000000..2b59a505136699 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_pipeline_tr.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Turkish mt5_small_3task_prepend_tquad2_pipeline pipeline T5Transformer from obss +author: John Snow Labs +name: mt5_small_3task_prepend_tquad2_pipeline +date: 2024-08-19 +tags: [tr, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: tr +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_3task_prepend_tquad2_pipeline` is a Turkish model originally trained by obss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_3task_prepend_tquad2_pipeline_tr_5.4.2_3.0_1724082743253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_3task_prepend_tquad2_pipeline_tr_5.4.2_3.0_1724082743253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_3task_prepend_tquad2_pipeline", lang = "tr") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_3task_prepend_tquad2_pipeline", lang = "tr") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_3task_prepend_tquad2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|tr| +|Size:|1.2 GB| + +## References + +https://huggingface.co/obss/mt5-small-3task-prepend-tquad2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_tr.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_tr.md new file mode 100644 index 00000000000000..fed23f4a5822d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_3task_prepend_tquad2_tr.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Turkish mt5_small_3task_prepend_tquad2 T5Transformer from obss +author: John Snow Labs +name: mt5_small_3task_prepend_tquad2 +date: 2024-08-19 +tags: [tr, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: tr +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_3task_prepend_tquad2` is a Turkish model originally trained by obss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_3task_prepend_tquad2_tr_5.4.2_3.0_1724082550330.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_3task_prepend_tquad2_tr_5.4.2_3.0_1724082550330.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_3task_prepend_tquad2","tr") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_3task_prepend_tquad2", "tr") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_3task_prepend_tquad2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|tr| +|Size:|1.2 GB| + +## References + +https://huggingface.co/obss/mt5-small-3task-prepend-tquad2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_en.md new file mode 100644 index 00000000000000..050b5b0a743c5a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ans_ext T5Transformer from sangrimlee +author: John Snow Labs +name: mt5_small_ans_ext +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ans_ext` is a English model originally trained by sangrimlee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ans_ext_en_5.4.2_3.0_1724064280431.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ans_ext_en_5.4.2_3.0_1724064280431.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ans_ext","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ans_ext", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ans_ext| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/sangrimlee/mt5-small-ans-ext \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_pipeline_en.md new file mode 100644 index 00000000000000..73813bd25bad7e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ans_ext_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ans_ext_pipeline pipeline T5Transformer from sangrimlee +author: John Snow Labs +name: mt5_small_ans_ext_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ans_ext_pipeline` is a English model originally trained by sangrimlee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ans_ext_pipeline_en_5.4.2_3.0_1724064469674.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ans_ext_pipeline_en_5.4.2_3.0_1724064469674.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ans_ext_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ans_ext_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ans_ext_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/sangrimlee/mt5-small-ans-ext + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_en.md new file mode 100644 index 00000000000000..a172aa88297ea8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_danish_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_danish_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_danish_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_danish_10k_en_5.4.2_3.0_1724075664950.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_danish_10k_en_5.4.2_3.0_1724075664950.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_danish_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_danish_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_danish_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-da-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_pipeline_en.md new file mode 100644 index 00000000000000..d81faf9faedef8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_danish_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_danish_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_danish_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_danish_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_danish_10k_pipeline_en_5.4.2_3.0_1724075825200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_danish_10k_pipeline_en_5.4.2_3.0_1724075825200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_danish_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_danish_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_danish_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-da-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_en.md new file mode 100644 index 00000000000000..6fde381a3c88b3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_eng_tata T5Transformer from adenhaus +author: John Snow Labs +name: mt5_small_eng_tata +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_eng_tata` is a English model originally trained by adenhaus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_eng_tata_en_5.4.2_3.0_1724109193287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_eng_tata_en_5.4.2_3.0_1724109193287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_eng_tata","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_eng_tata", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_eng_tata| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/adenhaus/mt5-small-eng-tata \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_pipeline_en.md new file mode 100644 index 00000000000000..bea187f849bee0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_eng_tata_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_eng_tata_pipeline pipeline T5Transformer from adenhaus +author: John Snow Labs +name: mt5_small_eng_tata_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_eng_tata_pipeline` is a English model originally trained by adenhaus. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_eng_tata_pipeline_en_5.4.2_3.0_1724109392985.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_eng_tata_pipeline_en_5.4.2_3.0_1724109392985.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_eng_tata_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_eng_tata_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_eng_tata_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/adenhaus/mt5-small-eng-tata + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_en.md new file mode 100644 index 00000000000000..f91b5503ead10a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_english_tonga_tonga_islands_hindi_akrov T5Transformer from AbhiKrov +author: John Snow Labs +name: mt5_small_english_tonga_tonga_islands_hindi_akrov +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_english_tonga_tonga_islands_hindi_akrov` is a English model originally trained by AbhiKrov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_english_tonga_tonga_islands_hindi_akrov_en_5.4.2_3.0_1724059240945.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_english_tonga_tonga_islands_hindi_akrov_en_5.4.2_3.0_1724059240945.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_english_tonga_tonga_islands_hindi_akrov","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_english_tonga_tonga_islands_hindi_akrov", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_english_tonga_tonga_islands_hindi_akrov| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AbhiKrov/mt5-small-english-to-hindi-akrov \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline_en.md new file mode 100644 index 00000000000000..3a9c0059269b68 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline pipeline T5Transformer from AbhiKrov +author: John Snow Labs +name: mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline` is a English model originally trained by AbhiKrov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline_en_5.4.2_3.0_1724059331672.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline_en_5.4.2_3.0_1724059331672.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_english_tonga_tonga_islands_hindi_akrov_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AbhiKrov/mt5-small-english-to-hindi-akrov + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_en.md new file mode 100644 index 00000000000000..58f3d8c706ff48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_13feb_5 T5Transformer from mqy +author: John Snow Labs +name: mt5_small_finetuned_13feb_5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_13feb_5` is a English model originally trained by mqy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_13feb_5_en_5.4.2_3.0_1724041571864.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_13feb_5_en_5.4.2_3.0_1724041571864.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_13feb_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_13feb_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_13feb_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/mqy/mt5-small-finetuned-13feb-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_pipeline_en.md new file mode 100644 index 00000000000000..68b1c844e568e6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_13feb_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_13feb_5_pipeline pipeline T5Transformer from mqy +author: John Snow Labs +name: mt5_small_finetuned_13feb_5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_13feb_5_pipeline` is a English model originally trained by mqy. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_13feb_5_pipeline_en_5.4.2_3.0_1724041758645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_13feb_5_pipeline_en_5.4.2_3.0_1724041758645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_13feb_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_13feb_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_13feb_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/mqy/mt5-small-finetuned-13feb-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_en.md new file mode 100644 index 00000000000000..d41b7c9c291b53 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_chinese_english_spanish T5Transformer from dfsj +author: John Snow Labs +name: mt5_small_finetuned_amazon_chinese_english_spanish +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_chinese_english_spanish` is a English model originally trained by dfsj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_english_spanish_en_5.4.2_3.0_1724040758808.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_english_spanish_en_5.4.2_3.0_1724040758808.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_chinese_english_spanish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_chinese_english_spanish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_chinese_english_spanish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/dfsj/mt5-small-finetuned-amazon-zh-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_pipeline_en.md new file mode 100644 index 00000000000000..b16f7e6ee50cf1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_chinese_english_spanish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_chinese_english_spanish_pipeline pipeline T5Transformer from dfsj +author: John Snow Labs +name: mt5_small_finetuned_amazon_chinese_english_spanish_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_chinese_english_spanish_pipeline` is a English model originally trained by dfsj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_english_spanish_pipeline_en_5.4.2_3.0_1724040847113.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_english_spanish_pipeline_en_5.4.2_3.0_1724040847113.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_chinese_english_spanish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_chinese_english_spanish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_chinese_english_spanish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/dfsj/mt5-small-finetuned-amazon-zh-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_en.md new file mode 100644 index 00000000000000..92a95b3b817130 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_anabach T5Transformer from AnaBach +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_anabach +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_anabach` is a English model originally trained by AnaBach. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_anabach_en_5.4.2_3.0_1724031550933.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_anabach_en_5.4.2_3.0_1724031550933.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_anabach","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_anabach", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_anabach| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AnaBach/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_pipeline_en.md new file mode 100644 index 00000000000000..579570fc47a4bc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_anabach_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_anabach_pipeline pipeline T5Transformer from AnaBach +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_anabach_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_anabach_pipeline` is a English model originally trained by AnaBach. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_anabach_pipeline_en_5.4.2_3.0_1724031643554.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_anabach_pipeline_en_5.4.2_3.0_1724031643554.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_anabach_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_anabach_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_anabach_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/AnaBach/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_en.md new file mode 100644 index 00000000000000..75c8bdf5819aae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_bobbyb1234 T5Transformer from BobbyB1234 +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_bobbyb1234 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_bobbyb1234` is a English model originally trained by BobbyB1234. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_bobbyb1234_en_5.4.2_3.0_1724037988536.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_bobbyb1234_en_5.4.2_3.0_1724037988536.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_bobbyb1234","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_bobbyb1234", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_bobbyb1234| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/BobbyB1234/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline_en.md new file mode 100644 index 00000000000000..19ee8c99b36a0d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline pipeline T5Transformer from BobbyB1234 +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline` is a English model originally trained by BobbyB1234. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline_en_5.4.2_3.0_1724038084291.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline_en_5.4.2_3.0_1724038084291.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_bobbyb1234_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/BobbyB1234/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_en.md new file mode 100644 index 00000000000000..b1aecd34c08015 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_cppmai T5Transformer from cppmai +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_cppmai +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_cppmai` is a English model originally trained by cppmai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cppmai_en_5.4.2_3.0_1724104235255.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cppmai_en_5.4.2_3.0_1724104235255.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_cppmai","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_cppmai", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_cppmai| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/cppmai/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline_en.md new file mode 100644 index 00000000000000..b03ff109e0be5b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline pipeline T5Transformer from cppmai +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline` is a English model originally trained by cppmai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline_en_5.4.2_3.0_1724104561836.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline_en_5.4.2_3.0_1724104561836.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_cppmai_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/cppmai/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_en.md new file mode 100644 index 00000000000000..c31769497b8b91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_elizabetvaganova T5Transformer from elizabetvaganova +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_elizabetvaganova +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_elizabetvaganova` is a English model originally trained by elizabetvaganova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_en_5.4.2_3.0_1724050216702.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_en_5.4.2_3.0_1724050216702.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_elizabetvaganova","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_elizabetvaganova", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_elizabetvaganova| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/elizabetvaganova/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline_en.md new file mode 100644 index 00000000000000..c677641feb700c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline pipeline T5Transformer from elizabetvaganova +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline` is a English model originally trained by elizabetvaganova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline_en_5.4.2_3.0_1724050329152.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline_en_5.4.2_3.0_1724050329152.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_elizabetvaganova_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/elizabetvaganova/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_en.md new file mode 100644 index 00000000000000..d028b101237721 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_katrinoushakova T5Transformer from katrinoushakova +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_katrinoushakova +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_katrinoushakova` is a English model originally trained by katrinoushakova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_katrinoushakova_en_5.4.2_3.0_1724064634695.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_katrinoushakova_en_5.4.2_3.0_1724064634695.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_katrinoushakova","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_katrinoushakova", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_katrinoushakova| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/katrinoushakova/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline_en.md new file mode 100644 index 00000000000000..d9f3ea49d166de --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline pipeline T5Transformer from katrinoushakova +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline` is a English model originally trained by katrinoushakova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline_en_5.4.2_3.0_1724064750565.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline_en_5.4.2_3.0_1724064750565.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_katrinoushakova_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/katrinoushakova/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_en.md new file mode 100644 index 00000000000000..1b134bae29b26f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_maydogdu T5Transformer from maydogdu +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_maydogdu +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_maydogdu` is a English model originally trained by maydogdu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_maydogdu_en_5.4.2_3.0_1724027561637.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_maydogdu_en_5.4.2_3.0_1724027561637.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_maydogdu","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_maydogdu", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_maydogdu| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/maydogdu/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline_en.md new file mode 100644 index 00000000000000..4aa4479dc3940a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline pipeline T5Transformer from maydogdu +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline` is a English model originally trained by maydogdu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline_en_5.4.2_3.0_1724027647610.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline_en_5.4.2_3.0_1724027647610.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_maydogdu_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/maydogdu/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_en.md new file mode 100644 index 00000000000000..b95f6da38397fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_stefanvrs T5Transformer from Stefanvrs +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_stefanvrs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_stefanvrs` is a English model originally trained by Stefanvrs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_stefanvrs_en_5.4.2_3.0_1724035748210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_stefanvrs_en_5.4.2_3.0_1724035748210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_stefanvrs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_stefanvrs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_stefanvrs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Stefanvrs/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline_en.md new file mode 100644 index 00000000000000..3f9c225a222ef0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline pipeline T5Transformer from Stefanvrs +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline` is a English model originally trained by Stefanvrs. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline_en_5.4.2_3.0_1724035851867.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline_en_5.4.2_3.0_1724035851867.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_stefanvrs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Stefanvrs/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_en.md new file mode 100644 index 00000000000000..19447b30ca1f1c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_japanese T5Transformer from eamar +author: John Snow Labs +name: mt5_small_finetuned_amazon_japanese +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_japanese` is a English model originally trained by eamar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_japanese_en_5.4.2_3.0_1724048592918.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_japanese_en_5.4.2_3.0_1724048592918.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_japanese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_japanese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_japanese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/eamar/mt5-small-finetuned-amazon-ja \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_pipeline_en.md new file mode 100644 index 00000000000000..7d83e6eae4d80a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_amazon_japanese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_japanese_pipeline pipeline T5Transformer from eamar +author: John Snow Labs +name: mt5_small_finetuned_amazon_japanese_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_japanese_pipeline` is a English model originally trained by eamar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_japanese_pipeline_en_5.4.2_3.0_1724048686886.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_japanese_pipeline_en_5.4.2_3.0_1724048686886.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_japanese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_japanese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_japanese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/eamar/mt5-small-finetuned-amazon-ja + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_en.md new file mode 100644 index 00000000000000..ab7e35ee703e26 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_billsum T5Transformer from kaiku03 +author: John Snow Labs +name: mt5_small_finetuned_billsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_billsum` is a English model originally trained by kaiku03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_billsum_en_5.4.2_3.0_1724076797130.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_billsum_en_5.4.2_3.0_1724076797130.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_billsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_billsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_billsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/kaiku03/mt5-small-finetuned-billsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_pipeline_en.md new file mode 100644 index 00000000000000..d20ea5715a013b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_billsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_billsum_pipeline pipeline T5Transformer from kaiku03 +author: John Snow Labs +name: mt5_small_finetuned_billsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_billsum_pipeline` is a English model originally trained by kaiku03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_billsum_pipeline_en_5.4.2_3.0_1724076896120.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_billsum_pipeline_en_5.4.2_3.0_1724076896120.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_billsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_billsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_billsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/kaiku03/mt5-small-finetuned-billsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline_xx.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline_xx.md new file mode 100644 index 00000000000000..9f2fd43734b972 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline_xx.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Multilingual mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline pipeline T5Transformer from CarryNid +author: John Snow Labs +name: mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline +date: 2024-08-19 +tags: [xx, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline` is a Multilingual model originally trained by CarryNid. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline_xx_5.4.2_3.0_1724082106512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline_xx_5.4.2_3.0_1724082106512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline", lang = "xx") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline", lang = "xx") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|xx| +|Size:|1.5 GB| + +## References + +https://huggingface.co/CarryNid/mt5-small-finetuned-multilingual-xlsum-new + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_xx.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_xx.md new file mode 100644 index 00000000000000..50745b687a7c09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_xx.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Multilingual mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid T5Transformer from CarryNid +author: John Snow Labs +name: mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid +date: 2024-08-19 +tags: [xx, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid` is a Multilingual model originally trained by CarryNid. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_xx_5.4.2_3.0_1724082027082.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid_xx_5.4.2_3.0_1724082027082.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid","xx") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid", "xx") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_multilingual_xlsum_nepal_bhasa_carrynid| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|xx| +|Size:|1.5 GB| + +## References + +https://huggingface.co/CarryNid/mt5-small-finetuned-multilingual-xlsum-new \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_en.md new file mode 100644 index 00000000000000..32c6f98eca148a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_sythetic T5Transformer from ignacioct +author: John Snow Labs +name: mt5_small_finetuned_sythetic +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_sythetic` is a English model originally trained by ignacioct. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_sythetic_en_5.4.2_3.0_1724030655012.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_sythetic_en_5.4.2_3.0_1724030655012.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_sythetic","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_sythetic", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_sythetic| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ignacioct/mt5-small-finetuned-sythetic \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_pipeline_en.md new file mode 100644 index 00000000000000..e27f7848739a13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_finetuned_sythetic_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_sythetic_pipeline pipeline T5Transformer from ignacioct +author: John Snow Labs +name: mt5_small_finetuned_sythetic_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_sythetic_pipeline` is a English model originally trained by ignacioct. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_sythetic_pipeline_en_5.4.2_3.0_1724030834660.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_sythetic_pipeline_en_5.4.2_3.0_1724030834660.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_sythetic_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_sythetic_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_sythetic_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ignacioct/mt5-small-finetuned-sythetic + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_en.md new file mode 100644 index 00000000000000..7a144afabdd543 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_frquad_qag_trimmed_50000 T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_frquad_qag_trimmed_50000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_frquad_qag_trimmed_50000` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qag_trimmed_50000_en_5.4.2_3.0_1724047076650.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qag_trimmed_50000_en_5.4.2_3.0_1724047076650.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_frquad_qag_trimmed_50000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_frquad_qag_trimmed_50000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_frquad_qag_trimmed_50000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|414.4 MB| + +## References + +https://huggingface.co/lmqg/mt5-small-frquad-qag-trimmed-50000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_pipeline_en.md new file mode 100644 index 00000000000000..027a2e91c234ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qag_trimmed_50000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_frquad_qag_trimmed_50000_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_frquad_qag_trimmed_50000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_frquad_qag_trimmed_50000_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qag_trimmed_50000_pipeline_en_5.4.2_3.0_1724047098100.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qag_trimmed_50000_pipeline_en_5.4.2_3.0_1724047098100.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_frquad_qag_trimmed_50000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_frquad_qag_trimmed_50000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_frquad_qag_trimmed_50000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|414.4 MB| + +## References + +https://huggingface.co/lmqg/mt5-small-frquad-qag-trimmed-50000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_en.md new file mode 100644 index 00000000000000..35b5153f442572 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_frquad_qg_trimmed_french_15000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_frquad_qg_trimmed_french_15000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_frquad_qg_trimmed_french_15000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qg_trimmed_french_15000_en_5.4.2_3.0_1724078425630.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qg_trimmed_french_15000_en_5.4.2_3.0_1724078425630.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_frquad_qg_trimmed_french_15000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_frquad_qg_trimmed_french_15000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_frquad_qg_trimmed_french_15000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|249.9 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-frquad-qg-trimmed-fr-15000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_pipeline_en.md new file mode 100644 index 00000000000000..a8295cc245c06b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_frquad_qg_trimmed_french_15000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_frquad_qg_trimmed_french_15000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_frquad_qg_trimmed_french_15000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_frquad_qg_trimmed_french_15000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qg_trimmed_french_15000_pipeline_en_5.4.2_3.0_1724078438452.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_frquad_qg_trimmed_french_15000_pipeline_en_5.4.2_3.0_1724078438452.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_frquad_qg_trimmed_french_15000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_frquad_qg_trimmed_french_15000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_frquad_qg_trimmed_french_15000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|249.9 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-frquad-qg-trimmed-fr-15000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_en.md new file mode 100644 index 00000000000000..49fa9964c30e63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_hebrew_news_oriya T5Transformer from virto +author: John Snow Labs +name: mt5_small_hebrew_news_oriya +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_hebrew_news_oriya` is a English model originally trained by virto. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_hebrew_news_oriya_en_5.4.2_3.0_1724037835134.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_hebrew_news_oriya_en_5.4.2_3.0_1724037835134.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_hebrew_news_oriya","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_hebrew_news_oriya", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_hebrew_news_oriya| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|4.6 MB| + +## References + +https://huggingface.co/virto/mt5-small-hebrew-news-or \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_pipeline_en.md new file mode 100644 index 00000000000000..57f91b1dd5e26f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_hebrew_news_oriya_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_hebrew_news_oriya_pipeline pipeline T5Transformer from virto +author: John Snow Labs +name: mt5_small_hebrew_news_oriya_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_hebrew_news_oriya_pipeline` is a English model originally trained by virto. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_hebrew_news_oriya_pipeline_en_5.4.2_3.0_1724037841099.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_hebrew_news_oriya_pipeline_en_5.4.2_3.0_1724037841099.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_hebrew_news_oriya_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_hebrew_news_oriya_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_hebrew_news_oriya_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|4.6 MB| + +## References + +https://huggingface.co/virto/mt5-small-hebrew-news-or + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_en.md new file mode 100644 index 00000000000000..258c4913aac4d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_itquad_qa_trimmed_italian_5000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_itquad_qa_trimmed_italian_5000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_itquad_qa_trimmed_italian_5000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_itquad_qa_trimmed_italian_5000_en_5.4.2_3.0_1724043992176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_itquad_qa_trimmed_italian_5000_en_5.4.2_3.0_1724043992176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_itquad_qa_trimmed_italian_5000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_itquad_qa_trimmed_italian_5000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_itquad_qa_trimmed_italian_5000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|196.1 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-itquad-qa-trimmed-it-5000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_pipeline_en.md new file mode 100644 index 00000000000000..f080e996f26669 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_itquad_qa_trimmed_italian_5000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_itquad_qa_trimmed_italian_5000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_itquad_qa_trimmed_italian_5000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_itquad_qa_trimmed_italian_5000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_itquad_qa_trimmed_italian_5000_pipeline_en_5.4.2_3.0_1724044001268.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_itquad_qa_trimmed_italian_5000_pipeline_en_5.4.2_3.0_1724044001268.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_itquad_qa_trimmed_italian_5000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_itquad_qa_trimmed_italian_5000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_itquad_qa_trimmed_italian_5000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|196.1 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-itquad-qa-trimmed-it-5000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_en.md new file mode 100644 index 00000000000000..6e2616477b6608 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_jaquad_qa_trimmed_japanese_15000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qa_trimmed_japanese_15000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qa_trimmed_japanese_15000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_15000_en_5.4.2_3.0_1724100227463.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_15000_en_5.4.2_3.0_1724100227463.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_jaquad_qa_trimmed_japanese_15000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_jaquad_qa_trimmed_japanese_15000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qa_trimmed_japanese_15000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|251.1 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qa-trimmed-ja-15000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline_en.md new file mode 100644 index 00000000000000..9941a4714ded13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline_en_5.4.2_3.0_1724100239822.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline_en_5.4.2_3.0_1724100239822.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qa_trimmed_japanese_15000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|251.1 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-jaquad-qa-trimmed-ja-15000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_ja.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_ja.md new file mode 100644 index 00000000000000..3c5898e44e8d62 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_ja.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Japanese mt5_small_jaquad_qag T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_jaquad_qag +date: 2024-08-19 +tags: [ja, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qag` is a Japanese model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qag_ja_5.4.2_3.0_1724078388054.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qag_ja_5.4.2_3.0_1724078388054.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_jaquad_qag","ja") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_jaquad_qag", "ja") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qag| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|1.2 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-jaquad-qag \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_pipeline_ja.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_pipeline_ja.md new file mode 100644 index 00000000000000..e614d6ec96de62 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_jaquad_qag_pipeline_ja.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Japanese mt5_small_jaquad_qag_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_jaquad_qag_pipeline +date: 2024-08-19 +tags: [ja, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_jaquad_qag_pipeline` is a Japanese model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qag_pipeline_ja_5.4.2_3.0_1724078476506.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_jaquad_qag_pipeline_ja_5.4.2_3.0_1724078476506.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_jaquad_qag_pipeline", lang = "ja") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_jaquad_qag_pipeline", lang = "ja") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_jaquad_qag_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ja| +|Size:|1.2 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-jaquad-qag + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_en.md new file mode 100644 index 00000000000000..2db9fc19d6d0b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_koquad_qa_trimmed_korean_5000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_koquad_qa_trimmed_korean_5000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_koquad_qa_trimmed_korean_5000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qa_trimmed_korean_5000_en_5.4.2_3.0_1724065571628.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qa_trimmed_korean_5000_en_5.4.2_3.0_1724065571628.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_koquad_qa_trimmed_korean_5000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_koquad_qa_trimmed_korean_5000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_koquad_qa_trimmed_korean_5000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|194.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-koquad-qa-trimmed-ko-5000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_pipeline_en.md new file mode 100644 index 00000000000000..6da9f4edb5bb75 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_koquad_qa_trimmed_korean_5000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_koquad_qa_trimmed_korean_5000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_koquad_qa_trimmed_korean_5000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_koquad_qa_trimmed_korean_5000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qa_trimmed_korean_5000_pipeline_en_5.4.2_3.0_1724065581532.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_koquad_qa_trimmed_korean_5000_pipeline_en_5.4.2_3.0_1724065581532.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_koquad_qa_trimmed_korean_5000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_koquad_qa_trimmed_korean_5000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_koquad_qa_trimmed_korean_5000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|194.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-koquad-qa-trimmed-ko-5000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_en.md new file mode 100644 index 00000000000000..00f63bc167c4e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_kyrgyz_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_kyrgyz_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_kyrgyz_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_kyrgyz_10k_en_5.4.2_3.0_1724029044707.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_kyrgyz_10k_en_5.4.2_3.0_1724029044707.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_kyrgyz_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_kyrgyz_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_kyrgyz_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ky-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_pipeline_en.md new file mode 100644 index 00000000000000..fcc1623266a5ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_kyrgyz_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_kyrgyz_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_kyrgyz_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_kyrgyz_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_kyrgyz_10k_pipeline_en_5.4.2_3.0_1724029210453.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_kyrgyz_10k_pipeline_en_5.4.2_3.0_1724029210453.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_kyrgyz_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_kyrgyz_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_kyrgyz_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ky-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_en.md new file mode 100644 index 00000000000000..d1446da9749cbe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_latvian_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_latvian_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_latvian_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_latvian_10k_en_5.4.2_3.0_1724109280104.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_latvian_10k_en_5.4.2_3.0_1724109280104.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_latvian_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_latvian_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_latvian_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-lv-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_pipeline_en.md new file mode 100644 index 00000000000000..1958620f50a5e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_latvian_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_latvian_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_latvian_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_latvian_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_latvian_10k_pipeline_en_5.4.2_3.0_1724109436013.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_latvian_10k_pipeline_en_5.4.2_3.0_1724109436013.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_latvian_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_latvian_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_latvian_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-lv-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_en.md new file mode 100644 index 00000000000000..c08f1983a9173f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_machinelearning T5Transformer from elizabetvaganova +author: John Snow Labs +name: mt5_small_machinelearning +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_machinelearning` is a English model originally trained by elizabetvaganova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_machinelearning_en_5.4.2_3.0_1724110276030.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_machinelearning_en_5.4.2_3.0_1724110276030.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_machinelearning","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_machinelearning", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_machinelearning| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/elizabetvaganova/mt5-small-machinelearning \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_pipeline_en.md new file mode 100644 index 00000000000000..ebf8532ae6f097 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_machinelearning_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_machinelearning_pipeline pipeline T5Transformer from elizabetvaganova +author: John Snow Labs +name: mt5_small_machinelearning_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_machinelearning_pipeline` is a English model originally trained by elizabetvaganova. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_machinelearning_pipeline_en_5.4.2_3.0_1724110398535.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_machinelearning_pipeline_en_5.4.2_3.0_1724110398535.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_machinelearning_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_machinelearning_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_machinelearning_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/elizabetvaganova/mt5-small-machinelearning + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_en.md new file mode 100644 index 00000000000000..f4d83602a12178 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_nc16_50k_ruen T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_50k_ruen +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_50k_ruen` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_50k_ruen_en_5.4.2_3.0_1724094884239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_50k_ruen_en_5.4.2_3.0_1724094884239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_nc16_50k_ruen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_nc16_50k_ruen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_50k_ruen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-50k-ruen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_pipeline_en.md new file mode 100644 index 00000000000000..bf3fa65f9222af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_nc16_50k_ruen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_nc16_50k_ruen_pipeline pipeline T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_50k_ruen_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_50k_ruen_pipeline` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_50k_ruen_pipeline_en_5.4.2_3.0_1724095054798.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_50k_ruen_pipeline_en_5.4.2_3.0_1724095054798.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_nc16_50k_ruen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_nc16_50k_ruen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_50k_ruen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-50k-ruen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_en.md new file mode 100644 index 00000000000000..cbf67555593d46 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ruquad_ae_trimmed_50000 T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_ruquad_ae_trimmed_50000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_ae_trimmed_50000` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_ae_trimmed_50000_en_5.4.2_3.0_1724040246747.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_ae_trimmed_50000_en_5.4.2_3.0_1724040246747.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ruquad_ae_trimmed_50000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ruquad_ae_trimmed_50000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_ae_trimmed_50000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|422.5 MB| + +## References + +https://huggingface.co/lmqg/mt5-small-ruquad-ae-trimmed-50000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_pipeline_en.md new file mode 100644 index 00000000000000..c8a4e070467a3d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_ae_trimmed_50000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ruquad_ae_trimmed_50000_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_ruquad_ae_trimmed_50000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_ae_trimmed_50000_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1724040267744.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_ae_trimmed_50000_pipeline_en_5.4.2_3.0_1724040267744.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ruquad_ae_trimmed_50000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ruquad_ae_trimmed_50000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_ae_trimmed_50000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|422.5 MB| + +## References + +https://huggingface.co/lmqg/mt5-small-ruquad-ae-trimmed-50000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_en.md new file mode 100644 index 00000000000000..1f7509f71fc823 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_10000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_10000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_10000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_10000_en_5.4.2_3.0_1724071909418.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_10000_en_5.4.2_3.0_1724071909418.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_10000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ruquad_qa_trimmed_russian_10000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_10000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|224.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-10000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_pipeline_en.md new file mode 100644 index 00000000000000..da9925b090008c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_ruquad_qa_trimmed_russian_10000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ruquad_qa_trimmed_russian_10000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_ruquad_qa_trimmed_russian_10000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ruquad_qa_trimmed_russian_10000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_10000_pipeline_en_5.4.2_3.0_1724071920776.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ruquad_qa_trimmed_russian_10000_pipeline_en_5.4.2_3.0_1724071920776.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_10000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ruquad_qa_trimmed_russian_10000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ruquad_qa_trimmed_russian_10000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|224.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-ruquad-qa-trimmed-ru-10000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_en.md new file mode 100644 index 00000000000000..4e701455fdefe5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_squad_qg T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_squad_qg +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qg` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qg_en_5.4.2_3.0_1724089899182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qg_en_5.4.2_3.0_1724089899182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_squad_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_squad_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-squad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_pipeline_en.md new file mode 100644 index 00000000000000..c6d69394133642 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_squad_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_squad_qg_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_small_squad_qg_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qg_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qg_pipeline_en_5.4.2_3.0_1724089986334.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qg_pipeline_en_5.4.2_3.0_1724089986334.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_squad_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_squad_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/lmqg/mt5-small-squad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_en.md new file mode 100644 index 00000000000000..d332e5030bc820 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_tatar_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_tatar_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_tatar_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_tatar_10k_en_5.4.2_3.0_1724066304917.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_tatar_10k_en_5.4.2_3.0_1724066304917.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_tatar_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_tatar_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_tatar_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-tt-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_pipeline_en.md new file mode 100644 index 00000000000000..15a302c65f306b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_tatar_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_tatar_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_tatar_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_tatar_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_tatar_10k_pipeline_en_5.4.2_3.0_1724066480383.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_tatar_10k_pipeline_en_5.4.2_3.0_1724066480383.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_tatar_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_tatar_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_tatar_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-tt-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_en.md new file mode 100644 index 00000000000000..5be80e80c435b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_telugu_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_telugu_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_telugu_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_telugu_10k_en_5.4.2_3.0_1724035077596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_telugu_10k_en_5.4.2_3.0_1724035077596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_telugu_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_telugu_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_telugu_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-te-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_pipeline_en.md new file mode 100644 index 00000000000000..583841bfcf6de4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_telugu_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_telugu_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_telugu_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_telugu_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_telugu_10k_pipeline_en_5.4.2_3.0_1724035234746.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_telugu_10k_pipeline_en_5.4.2_3.0_1724035234746.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_telugu_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_telugu_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_telugu_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-te-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_en.md new file mode 100644 index 00000000000000..06efae41bd2662 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_translation_thai_english T5Transformer from SuperAI2-Machima +author: John Snow Labs +name: mt5_small_translation_thai_english +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_translation_thai_english` is a English model originally trained by SuperAI2-Machima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_translation_thai_english_en_5.4.2_3.0_1724026805945.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_translation_thai_english_en_5.4.2_3.0_1724026805945.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_translation_thai_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_translation_thai_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_translation_thai_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/SuperAI2-Machima/mt5-small-translation_thai-english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_pipeline_en.md new file mode 100644 index 00000000000000..9ab30abb82d7a2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translation_thai_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_translation_thai_english_pipeline pipeline T5Transformer from SuperAI2-Machima +author: John Snow Labs +name: mt5_small_translation_thai_english_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_translation_thai_english_pipeline` is a English model originally trained by SuperAI2-Machima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_translation_thai_english_pipeline_en_5.4.2_3.0_1724026893439.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_translation_thai_english_pipeline_en_5.4.2_3.0_1724026893439.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_translation_thai_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_translation_thai_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_translation_thai_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/SuperAI2-Machima/mt5-small-translation_thai-english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_en.md new file mode 100644 index 00000000000000..aef82457856c16 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_translit T5Transformer from Saugatkafley +author: John Snow Labs +name: mt5_small_translit +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_translit` is a English model originally trained by Saugatkafley. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_translit_en_5.4.2_3.0_1724039766331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_translit_en_5.4.2_3.0_1724039766331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_translit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_translit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_translit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Saugatkafley/mt5-small-translit \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_pipeline_en.md new file mode 100644 index 00000000000000..2871ea9f2980bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_translit_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_translit_pipeline pipeline T5Transformer from Saugatkafley +author: John Snow Labs +name: mt5_small_translit_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_translit_pipeline` is a English model originally trained by Saugatkafley. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_translit_pipeline_en_5.4.2_3.0_1724039879926.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_translit_pipeline_en_5.4.2_3.0_1724039879926.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_translit_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_translit_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_translit_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Saugatkafley/mt5-small-translit + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_en.md new file mode 100644 index 00000000000000..b9fb582f7437ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_english_120000_squad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_120000_squad_qa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_120000_squad_qa` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_120000_squad_qa_en_5.4.2_3.0_1724082994439.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_120000_squad_qa_en_5.4.2_3.0_1724082994439.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_english_120000_squad_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_english_120000_squad_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_120000_squad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|743.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-120000-squad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_pipeline_en.md new file mode 100644 index 00000000000000..68870191f2ac9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_english_120000_squad_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_english_120000_squad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_english_120000_squad_qa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_english_120000_squad_qa_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_120000_squad_qa_pipeline_en_5.4.2_3.0_1724083039483.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_english_120000_squad_qa_pipeline_en_5.4.2_3.0_1724083039483.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_english_120000_squad_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_english_120000_squad_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_english_120000_squad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|743.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-en-120000-squad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_en.md new file mode 100644 index 00000000000000..f288692797eb96 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_italian_30000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_30000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_30000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_30000_en_5.4.2_3.0_1724049304728.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_30000_en_5.4.2_3.0_1724049304728.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_30000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_30000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_30000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|174.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-30000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_pipeline_en.md new file mode 100644 index 00000000000000..08fa6d0748d2db --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_30000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_italian_30000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_30000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_30000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_30000_pipeline_en_5.4.2_3.0_1724049363257.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_30000_pipeline_en_5.4.2_3.0_1724049363257.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_30000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_30000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_30000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|174.8 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-30000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_en.md new file mode 100644 index 00000000000000..774f1c0fac742b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_italian_60000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_60000 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_60000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_en_5.4.2_3.0_1724082412044.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_en_5.4.2_3.0_1724082412044.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_60000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_60000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_60000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|262.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-60000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_it.md new file mode 100644 index 00000000000000..4e1da109ce1212 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_it.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_60000_itquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_60000_itquad_qa +date: 2024-08-19 +tags: [it, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_60000_itquad_qa` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_itquad_qa_it_5.4.2_3.0_1724097579463.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_itquad_qa_it_5.4.2_3.0_1724097579463.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_60000_itquad_qa","it") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_60000_itquad_qa", "it") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_60000_itquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|it| +|Size:|483.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-60000-itquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_pipeline_it.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_pipeline_it.md new file mode 100644 index 00000000000000..b4f28f8dfbfc6e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_itquad_qa_pipeline_it.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_60000_itquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_60000_itquad_qa_pipeline +date: 2024-08-19 +tags: [it, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_60000_itquad_qa_pipeline` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_itquad_qa_pipeline_it_5.4.2_3.0_1724097605477.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_itquad_qa_pipeline_it_5.4.2_3.0_1724097605477.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_60000_itquad_qa_pipeline", lang = "it") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_60000_itquad_qa_pipeline", lang = "it") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_60000_itquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|it| +|Size:|483.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-60000-itquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_pipeline_en.md new file mode 100644 index 00000000000000..5f287215cf7024 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_italian_60000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_italian_60000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_60000_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_60000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_pipeline_en_5.4.2_3.0_1724082503156.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_60000_pipeline_en_5.4.2_3.0_1724082503156.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_60000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_60000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_60000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|262.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-60000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_ja.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_ja.md new file mode 100644 index 00000000000000..d6d7867d562508 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_ja.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Japanese mt5_small_trimmed_japanese_90000_jaquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_japanese_90000_jaquad_qa +date: 2024-08-19 +tags: [ja, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_japanese_90000_jaquad_qa` is a Japanese model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_90000_jaquad_qa_ja_5.4.2_3.0_1724085767080.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_90000_jaquad_qa_ja_5.4.2_3.0_1724085767080.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_japanese_90000_jaquad_qa","ja") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_japanese_90000_jaquad_qa", "ja") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_japanese_90000_jaquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ja| +|Size:|583.3 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ja-90000-jaquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline_ja.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline_ja.md new file mode 100644 index 00000000000000..708d89afe675d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline_ja.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Japanese mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline +date: 2024-08-19 +tags: [ja, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ja +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline` is a Japanese model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline_ja_5.4.2_3.0_1724085802385.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline_ja_5.4.2_3.0_1724085802385.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline", lang = "ja") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline", lang = "ja") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_japanese_90000_jaquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ja| +|Size:|583.3 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ja-90000-jaquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_pipeline_ru.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_pipeline_ru.md new file mode 100644 index 00000000000000..6b301ba2f895af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_pipeline_ru.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_120000_ruquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_120000_ruquad_qg_pipeline +date: 2024-08-19 +tags: [ru, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_120000_ruquad_qg_pipeline` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_120000_ruquad_qg_pipeline_ru_5.4.2_3.0_1724026709385.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_120000_ruquad_qg_pipeline_ru_5.4.2_3.0_1724026709385.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_russian_120000_ruquad_qg_pipeline", lang = "ru") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_russian_120000_ruquad_qg_pipeline", lang = "ru") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_120000_ruquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ru| +|Size:|705.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-120000-ruquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_ru.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_ru.md new file mode 100644 index 00000000000000..876df5780d9b4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_russian_120000_ruquad_qg_ru.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Russian mt5_small_trimmed_russian_120000_ruquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_russian_120000_ruquad_qg +date: 2024-08-19 +tags: [ru, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_russian_120000_ruquad_qg` is a Russian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_120000_ruquad_qg_ru_5.4.2_3.0_1724026664737.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_russian_120000_ruquad_qg_ru_5.4.2_3.0_1724026664737.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_120000_ruquad_qg","ru") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_russian_120000_ruquad_qg", "ru") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_russian_120000_ruquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ru| +|Size:|705.6 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ru-120000-ruquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_es.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_es.md new file mode 100644 index 00000000000000..a93437829cc39e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_10000_esquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_10000_esquad_qg +date: 2024-08-19 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_10000_esquad_qg` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_10000_esquad_qg_es_5.4.2_3.0_1724061683046.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_10000_esquad_qg_es_5.4.2_3.0_1724061683046.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_10000_esquad_qg","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_10000_esquad_qg", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_10000_esquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|224.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-10000-esquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_pipeline_es.md new file mode 100644 index 00000000000000..f1a9fad6253f18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_10000_esquad_qg_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_10000_esquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_10000_esquad_qg_pipeline +date: 2024-08-19 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_10000_esquad_qg_pipeline` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_10000_esquad_qg_pipeline_es_5.4.2_3.0_1724061694427.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_10000_esquad_qg_pipeline_es_5.4.2_3.0_1724061694427.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_spanish_10000_esquad_qg_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_spanish_10000_esquad_qg_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_10000_esquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|224.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-10000-esquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_es.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_es.md new file mode 100644 index 00000000000000..a5de22919c29e7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_es.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_90000_esquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_90000_esquad_qa +date: 2024-08-19 +tags: [es, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_90000_esquad_qa` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_90000_esquad_qa_es_5.4.2_3.0_1724069724520.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_90000_esquad_qa_es_5.4.2_3.0_1724069724520.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_90000_esquad_qa","es") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_spanish_90000_esquad_qa", "es") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_90000_esquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|es| +|Size:|617.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-90000-esquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_pipeline_es.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_pipeline_es.md new file mode 100644 index 00000000000000..c994cbd6e5659d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_trimmed_spanish_90000_esquad_qa_pipeline_es.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Castilian, Spanish mt5_small_trimmed_spanish_90000_esquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_spanish_90000_esquad_qa_pipeline +date: 2024-08-19 +tags: [es, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: es +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_spanish_90000_esquad_qa_pipeline` is a Castilian, Spanish model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_90000_esquad_qa_pipeline_es_5.4.2_3.0_1724069759929.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_spanish_90000_esquad_qa_pipeline_es_5.4.2_3.0_1724069759929.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_spanish_90000_esquad_qa_pipeline", lang = "es") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_spanish_90000_esquad_qa_pipeline", lang = "es") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_spanish_90000_esquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|es| +|Size:|617.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-es-90000-esquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_en.md new file mode 100644 index 00000000000000..7ffdbd20c38143 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_vietnamese_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_vietnamese_10k +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_vietnamese_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_vietnamese_10k_en_5.4.2_3.0_1724079873368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_vietnamese_10k_en_5.4.2_3.0_1724079873368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_vietnamese_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_vietnamese_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_vietnamese_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-vi-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_pipeline_en.md new file mode 100644 index 00000000000000..8f1ac9ccd7dd3b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_small_vietnamese_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_vietnamese_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_vietnamese_10k_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_vietnamese_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_vietnamese_10k_pipeline_en_5.4.2_3.0_1724080031349.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_vietnamese_10k_pipeline_en_5.4.2_3.0_1724080031349.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_vietnamese_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_vietnamese_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_vietnamese_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-vi-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_en.md new file mode 100644 index 00000000000000..c0729d34cdfdcc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_summarize_japanese_6051_japanese T5Transformer from ryusangwon +author: John Snow Labs +name: mt5_summarize_japanese_6051_japanese +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_summarize_japanese_6051_japanese` is a English model originally trained by ryusangwon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_summarize_japanese_6051_japanese_en_5.4.2_3.0_1724047166934.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_summarize_japanese_6051_japanese_en_5.4.2_3.0_1724047166934.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_summarize_japanese_6051_japanese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_summarize_japanese_6051_japanese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_summarize_japanese_6051_japanese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ryusangwon/mt5_summarize_japanese-6051-japanese \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_pipeline_en.md new file mode 100644 index 00000000000000..93a490aaf0a18e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_summarize_japanese_6051_japanese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_summarize_japanese_6051_japanese_pipeline pipeline T5Transformer from ryusangwon +author: John Snow Labs +name: mt5_summarize_japanese_6051_japanese_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_summarize_japanese_6051_japanese_pipeline` is a English model originally trained by ryusangwon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_summarize_japanese_6051_japanese_pipeline_en_5.4.2_3.0_1724047252822.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_summarize_japanese_6051_japanese_pipeline_en_5.4.2_3.0_1724047252822.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_summarize_japanese_6051_japanese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_summarize_japanese_6051_japanese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_summarize_japanese_6051_japanese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ryusangwon/mt5_summarize_japanese-6051-japanese + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_en.md new file mode 100644 index 00000000000000..d33c47527d6191 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_thai_sum_finetuned_sanook_news_headlines T5Transformer from sorayutmild +author: John Snow Labs +name: mt5_thai_sum_finetuned_sanook_news_headlines +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thai_sum_finetuned_sanook_news_headlines` is a English model originally trained by sorayutmild. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thai_sum_finetuned_sanook_news_headlines_en_5.4.2_3.0_1724092502287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thai_sum_finetuned_sanook_news_headlines_en_5.4.2_3.0_1724092502287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_thai_sum_finetuned_sanook_news_headlines","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_thai_sum_finetuned_sanook_news_headlines", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thai_sum_finetuned_sanook_news_headlines| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/sorayutmild/mt5-thai-sum-finetuned-sanook-news-headlines \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_pipeline_en.md new file mode 100644 index 00000000000000..61e13614f9945c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-mt5_thai_sum_finetuned_sanook_news_headlines_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_thai_sum_finetuned_sanook_news_headlines_pipeline pipeline T5Transformer from sorayutmild +author: John Snow Labs +name: mt5_thai_sum_finetuned_sanook_news_headlines_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_thai_sum_finetuned_sanook_news_headlines_pipeline` is a English model originally trained by sorayutmild. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_thai_sum_finetuned_sanook_news_headlines_pipeline_en_5.4.2_3.0_1724092663299.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_thai_sum_finetuned_sanook_news_headlines_pipeline_en_5.4.2_3.0_1724092663299.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_thai_sum_finetuned_sanook_news_headlines_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_thai_sum_finetuned_sanook_news_headlines_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_thai_sum_finetuned_sanook_news_headlines_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/sorayutmild/mt5-thai-sum-finetuned-sanook-news-headlines + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_en.md new file mode 100644 index 00000000000000..2ecf4923691bcb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English multi_news_model T5Transformer from Sif10 +author: John Snow Labs +name: multi_news_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`multi_news_model` is a English model originally trained by Sif10. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/multi_news_model_en_5.4.2_3.0_1724054048292.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/multi_news_model_en_5.4.2_3.0_1724054048292.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("multi_news_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("multi_news_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|multi_news_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Sif10/multi_news_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_pipeline_en.md new file mode 100644 index 00000000000000..b0a60b672a8d37 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-multi_news_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English multi_news_model_pipeline pipeline T5Transformer from Sif10 +author: John Snow Labs +name: multi_news_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`multi_news_model_pipeline` is a English model originally trained by Sif10. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/multi_news_model_pipeline_en_5.4.2_3.0_1724054099397.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/multi_news_model_pipeline_en_5.4.2_3.0_1724054099397.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("multi_news_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("multi_news_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|multi_news_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Sif10/multi_news_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_en.md b/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_en.md new file mode 100644 index 00000000000000..49288ac649b2a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English narrative_gen T5Transformer from we-are-groot +author: John Snow Labs +name: narrative_gen +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`narrative_gen` is a English model originally trained by we-are-groot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/narrative_gen_en_5.4.2_3.0_1724050290272.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/narrative_gen_en_5.4.2_3.0_1724050290272.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("narrative_gen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("narrative_gen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|narrative_gen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|287.5 MB| + +## References + +https://huggingface.co/we-are-groot/narrative_gen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_pipeline_en.md new file mode 100644 index 00000000000000..f88b9ce3fa7ffe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-narrative_gen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English narrative_gen_pipeline pipeline T5Transformer from we-are-groot +author: John Snow Labs +name: narrative_gen_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`narrative_gen_pipeline` is a English model originally trained by we-are-groot. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/narrative_gen_pipeline_en_5.4.2_3.0_1724050323442.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/narrative_gen_pipeline_en_5.4.2_3.0_1724050323442.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("narrative_gen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("narrative_gen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|narrative_gen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|287.5 MB| + +## References + +https://huggingface.co/we-are-groot/narrative_gen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-newssummarization_en.md b/docs/_posts/ahmedlone127/2024-08-19-newssummarization_en.md new file mode 100644 index 00000000000000..b1972022f7cc6a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-newssummarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English newssummarization T5Transformer from Ajeevan123 +author: John Snow Labs +name: newssummarization +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`newssummarization` is a English model originally trained by Ajeevan123. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/newssummarization_en_5.4.2_3.0_1724096116692.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/newssummarization_en_5.4.2_3.0_1724096116692.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("newssummarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("newssummarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|newssummarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/Ajeevan123/NewsSummarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-newssummarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-newssummarization_pipeline_en.md new file mode 100644 index 00000000000000..875ec7875eee61 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-newssummarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English newssummarization_pipeline pipeline T5Transformer from Ajeevan123 +author: John Snow Labs +name: newssummarization_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`newssummarization_pipeline` is a English model originally trained by Ajeevan123. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/newssummarization_pipeline_en_5.4.2_3.0_1724096139509.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/newssummarization_pipeline_en_5.4.2_3.0_1724096139509.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("newssummarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("newssummarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|newssummarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.2 MB| + +## References + +https://huggingface.co/Ajeevan123/NewsSummarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-newsumt5_en.md b/docs/_posts/ahmedlone127/2024-08-19-newsumt5_en.md new file mode 100644 index 00000000000000..87e3f178e5dfe9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-newsumt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English newsumt5 T5Transformer from Tawanmeansthesun +author: John Snow Labs +name: newsumt5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`newsumt5` is a English model originally trained by Tawanmeansthesun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/newsumt5_en_5.4.2_3.0_1724088069478.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/newsumt5_en_5.4.2_3.0_1724088069478.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("newsumt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("newsumt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|newsumt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|890.7 MB| + +## References + +https://huggingface.co/Tawanmeansthesun/newsumt5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-newsumt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-newsumt5_pipeline_en.md new file mode 100644 index 00000000000000..dc3be16f769afa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-newsumt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English newsumt5_pipeline pipeline T5Transformer from Tawanmeansthesun +author: John Snow Labs +name: newsumt5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`newsumt5_pipeline` is a English model originally trained by Tawanmeansthesun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/newsumt5_pipeline_en_5.4.2_3.0_1724088137292.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/newsumt5_pipeline_en_5.4.2_3.0_1724088137292.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("newsumt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("newsumt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|newsumt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|890.7 MB| + +## References + +https://huggingface.co/Tawanmeansthesun/newsumt5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_en.md b/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_en.md new file mode 100644 index 00000000000000..00a1f65f338dbf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nl2sql_cw T5Transformer from dhruvM +author: John Snow Labs +name: nl2sql_cw +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nl2sql_cw` is a English model originally trained by dhruvM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nl2sql_cw_en_5.4.2_3.0_1724037661324.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nl2sql_cw_en_5.4.2_3.0_1724037661324.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nl2sql_cw","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nl2sql_cw", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nl2sql_cw| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/dhruvM/NL2SQL-CW \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_pipeline_en.md new file mode 100644 index 00000000000000..393a9109d2ff58 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-nl2sql_cw_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nl2sql_cw_pipeline pipeline T5Transformer from dhruvM +author: John Snow Labs +name: nl2sql_cw_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nl2sql_cw_pipeline` is a English model originally trained by dhruvM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nl2sql_cw_pipeline_en_5.4.2_3.0_1724037793891.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nl2sql_cw_pipeline_en_5.4.2_3.0_1724037793891.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nl2sql_cw_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nl2sql_cw_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nl2sql_cw_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/dhruvM/NL2SQL-CW + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-output_aprab_en.md b/docs/_posts/ahmedlone127/2024-08-19-output_aprab_en.md new file mode 100644 index 00000000000000..a1db17daa04471 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-output_aprab_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English output_aprab T5Transformer from aprab +author: John Snow Labs +name: output_aprab +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`output_aprab` is a English model originally trained by aprab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/output_aprab_en_5.4.2_3.0_1724040990323.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/output_aprab_en_5.4.2_3.0_1724040990323.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("output_aprab","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("output_aprab", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|output_aprab| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.8 MB| + +## References + +https://huggingface.co/aprab/output \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-output_aprab_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-output_aprab_pipeline_en.md new file mode 100644 index 00000000000000..59c714b2e91f06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-output_aprab_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English output_aprab_pipeline pipeline T5Transformer from aprab +author: John Snow Labs +name: output_aprab_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`output_aprab_pipeline` is a English model originally trained by aprab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/output_aprab_pipeline_en_5.4.2_3.0_1724041010678.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/output_aprab_pipeline_en_5.4.2_3.0_1724041010678.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("output_aprab_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("output_aprab_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|output_aprab_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.8 MB| + +## References + +https://huggingface.co/aprab/output + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_en.md b/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_en.md new file mode 100644 index 00000000000000..096172ca66fe99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English p5_toys_small T5Transformer from makitanikaze +author: John Snow Labs +name: p5_toys_small +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`p5_toys_small` is a English model originally trained by makitanikaze. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/p5_toys_small_en_5.4.2_3.0_1724031452064.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/p5_toys_small_en_5.4.2_3.0_1724031452064.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("p5_toys_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("p5_toys_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|p5_toys_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.7 MB| + +## References + +https://huggingface.co/makitanikaze/P5_toys_small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_pipeline_en.md new file mode 100644 index 00000000000000..23df26c741e61f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-p5_toys_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English p5_toys_small_pipeline pipeline T5Transformer from makitanikaze +author: John Snow Labs +name: p5_toys_small_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`p5_toys_small_pipeline` is a English model originally trained by makitanikaze. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/p5_toys_small_pipeline_en_5.4.2_3.0_1724031468773.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/p5_toys_small_pipeline_en_5.4.2_3.0_1724031468773.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("p5_toys_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("p5_toys_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|p5_toys_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.7 MB| + +## References + +https://huggingface.co/makitanikaze/P5_toys_small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_en.md new file mode 100644 index 00000000000000..ecfd995302b5ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English pangea_summarization_model T5Transformer from ieuniversity +author: John Snow Labs +name: pangea_summarization_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pangea_summarization_model` is a English model originally trained by ieuniversity. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pangea_summarization_model_en_5.4.2_3.0_1724071923928.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pangea_summarization_model_en_5.4.2_3.0_1724071923928.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("pangea_summarization_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("pangea_summarization_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pangea_summarization_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|311.2 MB| + +## References + +https://huggingface.co/ieuniversity/pangea_summarization_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_pipeline_en.md new file mode 100644 index 00000000000000..4caeffebbb4a56 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pangea_summarization_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English pangea_summarization_model_pipeline pipeline T5Transformer from ieuniversity +author: John Snow Labs +name: pangea_summarization_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pangea_summarization_model_pipeline` is a English model originally trained by ieuniversity. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pangea_summarization_model_pipeline_en_5.4.2_3.0_1724071945856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pangea_summarization_model_pipeline_en_5.4.2_3.0_1724071945856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("pangea_summarization_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("pangea_summarization_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pangea_summarization_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|311.2 MB| + +## References + +https://huggingface.co/ieuniversity/pangea_summarization_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_en.md new file mode 100644 index 00000000000000..f4f6d308dcbfba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English paper_summarization T5Transformer from machinelearningzuu +author: John Snow Labs +name: paper_summarization +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`paper_summarization` is a English model originally trained by machinelearningzuu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/paper_summarization_en_5.4.2_3.0_1724082537637.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/paper_summarization_en_5.4.2_3.0_1724082537637.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("paper_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("paper_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|paper_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|300.2 MB| + +## References + +https://huggingface.co/machinelearningzuu/paper-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_pipeline_en.md new file mode 100644 index 00000000000000..a633a416ffea50 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-paper_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English paper_summarization_pipeline pipeline T5Transformer from machinelearningzuu +author: John Snow Labs +name: paper_summarization_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`paper_summarization_pipeline` is a English model originally trained by machinelearningzuu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/paper_summarization_pipeline_en_5.4.2_3.0_1724082559331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/paper_summarization_pipeline_en_5.4.2_3.0_1724082559331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("paper_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("paper_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|paper_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|300.2 MB| + +## References + +https://huggingface.co/machinelearningzuu/paper-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_en.md b/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_en.md new file mode 100644 index 00000000000000..8c4c2405961048 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English pima_diabetes T5Transformer from ahsan-mavros +author: John Snow Labs +name: pima_diabetes +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pima_diabetes` is a English model originally trained by ahsan-mavros. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pima_diabetes_en_5.4.2_3.0_1724070763286.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pima_diabetes_en_5.4.2_3.0_1724070763286.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("pima_diabetes","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("pima_diabetes", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pima_diabetes| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ahsan-mavros/pima-diabetes \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_pipeline_en.md new file mode 100644 index 00000000000000..0bf024bde7735a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pima_diabetes_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English pima_diabetes_pipeline pipeline T5Transformer from ahsan-mavros +author: John Snow Labs +name: pima_diabetes_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pima_diabetes_pipeline` is a English model originally trained by ahsan-mavros. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pima_diabetes_pipeline_en_5.4.2_3.0_1724070811097.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pima_diabetes_pipeline_en_5.4.2_3.0_1724070811097.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("pima_diabetes_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("pima_diabetes_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pima_diabetes_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ahsan-mavros/pima-diabetes + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-play1_en.md b/docs/_posts/ahmedlone127/2024-08-19-play1_en.md new file mode 100644 index 00000000000000..9633fd40f73e0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-play1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English play1 T5Transformer from RAYZ +author: John Snow Labs +name: play1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`play1` is a English model originally trained by RAYZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/play1_en_5.4.2_3.0_1724075016435.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/play1_en_5.4.2_3.0_1724075016435.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("play1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("play1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|play1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RAYZ/play1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-play1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-play1_pipeline_en.md new file mode 100644 index 00000000000000..f20e7de6bc03f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-play1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English play1_pipeline pipeline T5Transformer from RAYZ +author: John Snow Labs +name: play1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`play1_pipeline` is a English model originally trained by RAYZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/play1_pipeline_en_5.4.2_3.0_1724075065781.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/play1_pipeline_en_5.4.2_3.0_1724075065781.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("play1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("play1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|play1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RAYZ/play1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-play2_en.md b/docs/_posts/ahmedlone127/2024-08-19-play2_en.md new file mode 100644 index 00000000000000..5e59e69eec5c8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-play2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English play2 T5Transformer from RAYZ +author: John Snow Labs +name: play2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`play2` is a English model originally trained by RAYZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/play2_en_5.4.2_3.0_1724096751779.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/play2_en_5.4.2_3.0_1724096751779.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("play2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("play2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|play2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RAYZ/play2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-play2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-play2_pipeline_en.md new file mode 100644 index 00000000000000..38029f43321e3d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-play2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English play2_pipeline pipeline T5Transformer from RAYZ +author: John Snow Labs +name: play2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`play2_pipeline` is a English model originally trained by RAYZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/play2_pipeline_en_5.4.2_3.0_1724096802500.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/play2_pipeline_en_5.4.2_3.0_1724096802500.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("play2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("play2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|play2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RAYZ/play2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_en.md new file mode 100644 index 00000000000000..2b1b57baca330f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English plt5_base_poquad_qa_v2 T5Transformer from clarin-knext +author: John Snow Labs +name: plt5_base_poquad_qa_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`plt5_base_poquad_qa_v2` is a English model originally trained by clarin-knext. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/plt5_base_poquad_qa_v2_en_5.4.2_3.0_1724099087287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/plt5_base_poquad_qa_v2_en_5.4.2_3.0_1724099087287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("plt5_base_poquad_qa_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("plt5_base_poquad_qa_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|plt5_base_poquad_qa_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/clarin-knext/plt5-base-poquad-qa-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_pipeline_en.md new file mode 100644 index 00000000000000..e426b56e2b4cf6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-plt5_base_poquad_qa_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English plt5_base_poquad_qa_v2_pipeline pipeline T5Transformer from clarin-knext +author: John Snow Labs +name: plt5_base_poquad_qa_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`plt5_base_poquad_qa_v2_pipeline` is a English model originally trained by clarin-knext. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/plt5_base_poquad_qa_v2_pipeline_en_5.4.2_3.0_1724099155867.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/plt5_base_poquad_qa_v2_pipeline_en_5.4.2_3.0_1724099155867.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("plt5_base_poquad_qa_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("plt5_base_poquad_qa_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|plt5_base_poquad_qa_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/clarin-knext/plt5-base-poquad-qa-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_en.md b/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_en.md new file mode 100644 index 00000000000000..636fc175455d65 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English poem_gen_t5_small_v1 T5Transformer from DrishtiSharma +author: John Snow Labs +name: poem_gen_t5_small_v1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poem_gen_t5_small_v1` is a English model originally trained by DrishtiSharma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_v1_en_5.4.2_3.0_1724051733084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_v1_en_5.4.2_3.0_1724051733084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("poem_gen_t5_small_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("poem_gen_t5_small_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poem_gen_t5_small_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.9 MB| + +## References + +https://huggingface.co/DrishtiSharma/poem-gen-t5-small_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_pipeline_en.md new file mode 100644 index 00000000000000..1cc95c0f2c142c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-poem_gen_t5_small_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English poem_gen_t5_small_v1_pipeline pipeline T5Transformer from DrishtiSharma +author: John Snow Labs +name: poem_gen_t5_small_v1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`poem_gen_t5_small_v1_pipeline` is a English model originally trained by DrishtiSharma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_v1_pipeline_en_5.4.2_3.0_1724051752412.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/poem_gen_t5_small_v1_pipeline_en_5.4.2_3.0_1724051752412.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("poem_gen_t5_small_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("poem_gen_t5_small_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|poem_gen_t5_small_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.9 MB| + +## References + +https://huggingface.co/DrishtiSharma/poem-gen-t5-small_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_en.md b/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_en.md new file mode 100644 index 00000000000000..d211d7f2c36dff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English polish_transliterator_test1 T5Transformer from marcus2000 +author: John Snow Labs +name: polish_transliterator_test1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polish_transliterator_test1` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polish_transliterator_test1_en_5.4.2_3.0_1724031149888.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polish_transliterator_test1_en_5.4.2_3.0_1724031149888.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("polish_transliterator_test1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("polish_transliterator_test1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polish_transliterator_test1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|924.5 MB| + +## References + +https://huggingface.co/marcus2000/polish_transliterator_test1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_pipeline_en.md new file mode 100644 index 00000000000000..bfa17f02d82354 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-polish_transliterator_test1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English polish_transliterator_test1_pipeline pipeline T5Transformer from marcus2000 +author: John Snow Labs +name: polish_transliterator_test1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`polish_transliterator_test1_pipeline` is a English model originally trained by marcus2000. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/polish_transliterator_test1_pipeline_en_5.4.2_3.0_1724031233407.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/polish_transliterator_test1_pipeline_en_5.4.2_3.0_1724031233407.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("polish_transliterator_test1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("polish_transliterator_test1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|polish_transliterator_test1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|924.5 MB| + +## References + +https://huggingface.co/marcus2000/polish_transliterator_test1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_en.md b/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_en.md new file mode 100644 index 00000000000000..b015a26a14c235 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English pretrained_arat5_unlabeled T5Transformer from imen11111 +author: John Snow Labs +name: pretrained_arat5_unlabeled +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pretrained_arat5_unlabeled` is a English model originally trained by imen11111. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pretrained_arat5_unlabeled_en_5.4.2_3.0_1724027768117.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pretrained_arat5_unlabeled_en_5.4.2_3.0_1724027768117.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("pretrained_arat5_unlabeled","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("pretrained_arat5_unlabeled", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pretrained_arat5_unlabeled| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/imen11111/Pretrained_araT5_unlabeled \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_pipeline_en.md new file mode 100644 index 00000000000000..abb54fe67fd38c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pretrained_arat5_unlabeled_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English pretrained_arat5_unlabeled_pipeline pipeline T5Transformer from imen11111 +author: John Snow Labs +name: pretrained_arat5_unlabeled_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pretrained_arat5_unlabeled_pipeline` is a English model originally trained by imen11111. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pretrained_arat5_unlabeled_pipeline_en_5.4.2_3.0_1724027846075.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pretrained_arat5_unlabeled_pipeline_en_5.4.2_3.0_1724027846075.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("pretrained_arat5_unlabeled_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("pretrained_arat5_unlabeled_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pretrained_arat5_unlabeled_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/imen11111/Pretrained_araT5_unlabeled + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_en.md b/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_en.md new file mode 100644 index 00000000000000..6e900facef2934 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ptt5_xlsumm_30epochs T5Transformer from arthurmluz +author: John Snow Labs +name: ptt5_xlsumm_30epochs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_xlsumm_30epochs` is a English model originally trained by arthurmluz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_xlsumm_30epochs_en_5.4.2_3.0_1724082226322.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_xlsumm_30epochs_en_5.4.2_3.0_1724082226322.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ptt5_xlsumm_30epochs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ptt5_xlsumm_30epochs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_xlsumm_30epochs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|993.2 MB| + +## References + +https://huggingface.co/arthurmluz/ptt5-xlsumm-30epochs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_pipeline_en.md new file mode 100644 index 00000000000000..915a0997d195e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ptt5_xlsumm_30epochs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ptt5_xlsumm_30epochs_pipeline pipeline T5Transformer from arthurmluz +author: John Snow Labs +name: ptt5_xlsumm_30epochs_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ptt5_xlsumm_30epochs_pipeline` is a English model originally trained by arthurmluz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ptt5_xlsumm_30epochs_pipeline_en_5.4.2_3.0_1724082288045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ptt5_xlsumm_30epochs_pipeline_en_5.4.2_3.0_1724082288045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ptt5_xlsumm_30epochs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ptt5_xlsumm_30epochs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ptt5_xlsumm_30epochs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|993.2 MB| + +## References + +https://huggingface.co/arthurmluz/ptt5-xlsumm-30epochs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_en.md new file mode 100644 index 00000000000000..5bd662596c872c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English punctuation_nilc_t5_base T5Transformer from tiagoblima +author: John Snow Labs +name: punctuation_nilc_t5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`punctuation_nilc_t5_base` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/punctuation_nilc_t5_base_en_5.4.2_3.0_1724072014806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/punctuation_nilc_t5_base_en_5.4.2_3.0_1724072014806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("punctuation_nilc_t5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("punctuation_nilc_t5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|punctuation_nilc_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|968.5 MB| + +## References + +https://huggingface.co/tiagoblima/punctuation-nilc-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_pipeline_en.md new file mode 100644 index 00000000000000..04c7f8133a4c73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-punctuation_nilc_t5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English punctuation_nilc_t5_base_pipeline pipeline T5Transformer from tiagoblima +author: John Snow Labs +name: punctuation_nilc_t5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`punctuation_nilc_t5_base_pipeline` is a English model originally trained by tiagoblima. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/punctuation_nilc_t5_base_pipeline_en_5.4.2_3.0_1724072076210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/punctuation_nilc_t5_base_pipeline_en_5.4.2_3.0_1724072076210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("punctuation_nilc_t5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("punctuation_nilc_t5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|punctuation_nilc_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|968.6 MB| + +## References + +https://huggingface.co/tiagoblima/punctuation-nilc-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline_xx.md b/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline_xx.md new file mode 100644 index 00000000000000..d10b922513ebbe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline_xx.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Multilingual pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline pipeline T5Transformer from amir22010 +author: John Snow Labs +name: pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline +date: 2024-08-19 +tags: [xx, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline` is a Multilingual model originally trained by amir22010. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline_xx_5.4.2_3.0_1724053466432.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline_xx_5.4.2_3.0_1724053466432.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline", lang = "xx") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline", lang = "xx") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|xx| +|Size:|936.9 MB| + +## References + +https://huggingface.co/amir22010/PyABSA_Hospital_Multilingual_allenai_tk-instruct-base-def-pos_FinedTuned_Model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_xx.md b/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_xx.md new file mode 100644 index 00000000000000..f35e7d9627fdba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_xx.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Multilingual pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model T5Transformer from amir22010 +author: John Snow Labs +name: pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model +date: 2024-08-19 +tags: [xx, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model` is a Multilingual model originally trained by amir22010. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_xx_5.4.2_3.0_1724053415618.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model_xx_5.4.2_3.0_1724053415618.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model","xx") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model", "xx") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pyabsa_hospital_multilingual_allenai_turkmen_instruct_base_def_sayula_popoluca_finedtuned_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|xx| +|Size:|936.9 MB| + +## References + +https://huggingface.co/amir22010/PyABSA_Hospital_Multilingual_allenai_tk-instruct-base-def-pos_FinedTuned_Model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_en.md b/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_en.md new file mode 100644 index 00000000000000..bcccd8d6b131fe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qa_generator_varosa T5Transformer from Varosa +author: John Snow Labs +name: qa_generator_varosa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qa_generator_varosa` is a English model originally trained by Varosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qa_generator_varosa_en_5.4.2_3.0_1724050345423.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qa_generator_varosa_en_5.4.2_3.0_1724050345423.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qa_generator_varosa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qa_generator_varosa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qa_generator_varosa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Varosa/qa-generator \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_pipeline_en.md new file mode 100644 index 00000000000000..c8fb1bb7ca7879 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qa_generator_varosa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qa_generator_varosa_pipeline pipeline T5Transformer from Varosa +author: John Snow Labs +name: qa_generator_varosa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qa_generator_varosa_pipeline` is a English model originally trained by Varosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qa_generator_varosa_pipeline_en_5.4.2_3.0_1724050484158.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qa_generator_varosa_pipeline_en_5.4.2_3.0_1724050484158.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qa_generator_varosa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qa_generator_varosa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qa_generator_varosa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Varosa/qa-generator + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_en.md b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_en.md new file mode 100644 index 00000000000000..488f5a3e4c84ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qcpg_parabk2_sbert_lr1e_4 T5Transformer from madhavsankar +author: John Snow Labs +name: qcpg_parabk2_sbert_lr1e_4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_parabk2_sbert_lr1e_4` is a English model originally trained by madhavsankar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_sbert_lr1e_4_en_5.4.2_3.0_1724068426601.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_sbert_lr1e_4_en_5.4.2_3.0_1724068426601.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qcpg_parabk2_sbert_lr1e_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qcpg_parabk2_sbert_lr1e_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_parabk2_sbert_lr1e_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/madhavsankar/qcpg-parabk2-sbert-lr1e-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_pipeline_en.md new file mode 100644 index 00000000000000..29f10307e09ae8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_sbert_lr1e_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qcpg_parabk2_sbert_lr1e_4_pipeline pipeline T5Transformer from madhavsankar +author: John Snow Labs +name: qcpg_parabk2_sbert_lr1e_4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_parabk2_sbert_lr1e_4_pipeline` is a English model originally trained by madhavsankar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_sbert_lr1e_4_pipeline_en_5.4.2_3.0_1724068473912.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_sbert_lr1e_4_pipeline_en_5.4.2_3.0_1724068473912.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qcpg_parabk2_sbert_lr1e_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qcpg_parabk2_sbert_lr1e_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_parabk2_sbert_lr1e_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/madhavsankar/qcpg-parabk2-sbert-lr1e-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_en.md new file mode 100644 index 00000000000000..8d78ebd70b2cbd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qcpg_parabk2_t5_base T5Transformer from bibekyess +author: John Snow Labs +name: qcpg_parabk2_t5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_parabk2_t5_base` is a English model originally trained by bibekyess. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_t5_base_en_5.4.2_3.0_1724111125109.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_t5_base_en_5.4.2_3.0_1724111125109.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qcpg_parabk2_t5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qcpg_parabk2_t5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_parabk2_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bibekyess/qcpg-parabk2-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_pipeline_en.md new file mode 100644 index 00000000000000..74a978ab30a337 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qcpg_parabk2_t5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qcpg_parabk2_t5_base_pipeline pipeline T5Transformer from bibekyess +author: John Snow Labs +name: qcpg_parabk2_t5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qcpg_parabk2_t5_base_pipeline` is a English model originally trained by bibekyess. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_t5_base_pipeline_en_5.4.2_3.0_1724111173292.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qcpg_parabk2_t5_base_pipeline_en_5.4.2_3.0_1724111173292.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qcpg_parabk2_t5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qcpg_parabk2_t5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qcpg_parabk2_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bibekyess/qcpg-parabk2-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_en.md b/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_en.md new file mode 100644 index 00000000000000..6bdaf63c131d6a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qnli_t5_base_seed_3 T5Transformer from utahnlp +author: John Snow Labs +name: qnli_t5_base_seed_3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qnli_t5_base_seed_3` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qnli_t5_base_seed_3_en_5.4.2_3.0_1724069354087.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qnli_t5_base_seed_3_en_5.4.2_3.0_1724069354087.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qnli_t5_base_seed_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qnli_t5_base_seed_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qnli_t5_base_seed_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|980.9 MB| + +## References + +https://huggingface.co/utahnlp/qnli_t5-base_seed-3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_pipeline_en.md new file mode 100644 index 00000000000000..b2650c166b6295 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-qnli_t5_base_seed_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qnli_t5_base_seed_3_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: qnli_t5_base_seed_3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qnli_t5_base_seed_3_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qnli_t5_base_seed_3_pipeline_en_5.4.2_3.0_1724069411293.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qnli_t5_base_seed_3_pipeline_en_5.4.2_3.0_1724069411293.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qnli_t5_base_seed_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qnli_t5_base_seed_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qnli_t5_base_seed_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|980.9 MB| + +## References + +https://huggingface.co/utahnlp/qnli_t5-base_seed-3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ques_gen_en.md b/docs/_posts/ahmedlone127/2024-08-19-ques_gen_en.md new file mode 100644 index 00000000000000..32bb3436f6f903 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ques_gen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ques_gen T5Transformer from santhoshkolloju +author: John Snow Labs +name: ques_gen +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ques_gen` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ques_gen_en_5.4.2_3.0_1724100795121.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ques_gen_en_5.4.2_3.0_1724100795121.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ques_gen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ques_gen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ques_gen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/santhoshkolloju/ques_gen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ques_gen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-ques_gen_pipeline_en.md new file mode 100644 index 00000000000000..8d6d1421a8a4ee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ques_gen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ques_gen_pipeline pipeline T5Transformer from santhoshkolloju +author: John Snow Labs +name: ques_gen_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ques_gen_pipeline` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ques_gen_pipeline_en_5.4.2_3.0_1724100843873.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ques_gen_pipeline_en_5.4.2_3.0_1724100843873.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ques_gen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ques_gen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ques_gen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/santhoshkolloju/ques_gen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_en.md b/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_en.md new file mode 100644 index 00000000000000..e371147ee56ee5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rankt5_large T5Transformer from Soyoung97 +author: John Snow Labs +name: rankt5_large +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rankt5_large` is a English model originally trained by Soyoung97. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rankt5_large_en_5.4.2_3.0_1724049705045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rankt5_large_en_5.4.2_3.0_1724049705045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rankt5_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rankt5_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rankt5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/Soyoung97/RankT5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_pipeline_en.md new file mode 100644 index 00000000000000..7f397aed5b0b67 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rankt5_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rankt5_large_pipeline pipeline T5Transformer from Soyoung97 +author: John Snow Labs +name: rankt5_large_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rankt5_large_pipeline` is a English model originally trained by Soyoung97. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rankt5_large_pipeline_en_5.4.2_3.0_1724049869368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rankt5_large_pipeline_en_5.4.2_3.0_1724049869368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rankt5_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rankt5_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rankt5_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/Soyoung97/RankT5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_en.md new file mode 100644 index 00000000000000..703b593e269422 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English results_ara_t5_v2 T5Transformer from hiba2 +author: John Snow Labs +name: results_ara_t5_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_ara_t5_v2` is a English model originally trained by hiba2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_ara_t5_v2_en_5.4.2_3.0_1724029737745.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_ara_t5_v2_en_5.4.2_3.0_1724029737745.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("results_ara_t5_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("results_ara_t5_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_ara_t5_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|971.8 MB| + +## References + +https://huggingface.co/hiba2/results_ara_t5_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_pipeline_en.md new file mode 100644 index 00000000000000..cc7f0f112208ad --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-results_ara_t5_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English results_ara_t5_v2_pipeline pipeline T5Transformer from hiba2 +author: John Snow Labs +name: results_ara_t5_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`results_ara_t5_v2_pipeline` is a English model originally trained by hiba2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/results_ara_t5_v2_pipeline_en_5.4.2_3.0_1724029805414.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/results_ara_t5_v2_pipeline_en_5.4.2_3.0_1724029805414.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("results_ara_t5_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("results_ara_t5_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|results_ara_t5_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|971.8 MB| + +## References + +https://huggingface.co/hiba2/results_ara_t5_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_en.md new file mode 100644 index 00000000000000..fb0b7fb58935ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rg_model_base T5Transformer from songbo +author: John Snow Labs +name: rg_model_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rg_model_base` is a English model originally trained by songbo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rg_model_base_en_5.4.2_3.0_1724077211900.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rg_model_base_en_5.4.2_3.0_1724077211900.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rg_model_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rg_model_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rg_model_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/songbo/rg_model_base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_pipeline_en.md new file mode 100644 index 00000000000000..f776196e0d6205 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rg_model_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rg_model_base_pipeline pipeline T5Transformer from songbo +author: John Snow Labs +name: rg_model_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rg_model_base_pipeline` is a English model originally trained by songbo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rg_model_base_pipeline_en_5.4.2_3.0_1724077259703.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rg_model_base_pipeline_en_5.4.2_3.0_1724077259703.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rg_model_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rg_model_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rg_model_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/songbo/rg_model_base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_en.md new file mode 100644 index 00000000000000..144bdb6642c7b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rotten_tomatoes_t5_large_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: rotten_tomatoes_t5_large_seed_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rotten_tomatoes_t5_large_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rotten_tomatoes_t5_large_seed_1_en_5.4.2_3.0_1724085603939.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rotten_tomatoes_t5_large_seed_1_en_5.4.2_3.0_1724085603939.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rotten_tomatoes_t5_large_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rotten_tomatoes_t5_large_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rotten_tomatoes_t5_large_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/utahnlp/rotten_tomatoes_t5-large_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..351e6412584375 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rotten_tomatoes_t5_large_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rotten_tomatoes_t5_large_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: rotten_tomatoes_t5_large_seed_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rotten_tomatoes_t5_large_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rotten_tomatoes_t5_large_seed_1_pipeline_en_5.4.2_3.0_1724085775595.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rotten_tomatoes_t5_large_seed_1_pipeline_en_5.4.2_3.0_1724085775595.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rotten_tomatoes_t5_large_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rotten_tomatoes_t5_large_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rotten_tomatoes_t5_large_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/utahnlp/rotten_tomatoes_t5-large_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_en.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_en.md new file mode 100644 index 00000000000000..44500c81ffe390 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rut5_base_review T5Transformer from cointegrated +author: John Snow Labs +name: rut5_base_review +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_review` is a English model originally trained by cointegrated. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_review_en_5.4.2_3.0_1724050555629.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_review_en_5.4.2_3.0_1724050555629.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_base_review","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_base_review", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_review| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|996.5 MB| + +## References + +https://huggingface.co/cointegrated/rut5-base-review \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_pipeline_en.md new file mode 100644 index 00000000000000..c9579db4f5b977 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_review_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rut5_base_review_pipeline pipeline T5Transformer from cointegrated +author: John Snow Labs +name: rut5_base_review_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_review_pipeline` is a English model originally trained by cointegrated. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_review_pipeline_en_5.4.2_3.0_1724050603571.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_review_pipeline_en_5.4.2_3.0_1724050603571.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_base_review_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_base_review_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_review_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|996.5 MB| + +## References + +https://huggingface.co/cointegrated/rut5-base-review + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_en.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_en.md new file mode 100644 index 00000000000000..0250bda9bd32ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English rut5_base_texificator T5Transformer from basic-go +author: John Snow Labs +name: rut5_base_texificator +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_texificator` is a English model originally trained by basic-go. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_texificator_en_5.4.2_3.0_1724033151796.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_texificator_en_5.4.2_3.0_1724033151796.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_base_texificator","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_base_texificator", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_texificator| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|947.4 MB| + +## References + +https://huggingface.co/basic-go/rut5-base-texificator \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_pipeline_en.md new file mode 100644 index 00000000000000..43ce2beb3086bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_base_texificator_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English rut5_base_texificator_pipeline pipeline T5Transformer from basic-go +author: John Snow Labs +name: rut5_base_texificator_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_base_texificator_pipeline` is a English model originally trained by basic-go. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_base_texificator_pipeline_en_5.4.2_3.0_1724033206118.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_base_texificator_pipeline_en_5.4.2_3.0_1724033206118.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_base_texificator_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_base_texificator_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_base_texificator_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|947.4 MB| + +## References + +https://huggingface.co/basic-go/rut5-base-texificator + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_pipeline_ru.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_pipeline_ru.md new file mode 100644 index 00000000000000..d71d0a27e5687e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_pipeline_ru.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Russian rut5_large_sum_gazeta_pipeline pipeline T5Transformer from mlenjoyneer +author: John Snow Labs +name: rut5_large_sum_gazeta_pipeline +date: 2024-08-19 +tags: [ru, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_large_sum_gazeta_pipeline` is a Russian model originally trained by mlenjoyneer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_large_sum_gazeta_pipeline_ru_5.4.2_3.0_1724060144921.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_large_sum_gazeta_pipeline_ru_5.4.2_3.0_1724060144921.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_large_sum_gazeta_pipeline", lang = "ru") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_large_sum_gazeta_pipeline", lang = "ru") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_large_sum_gazeta_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ru| +|Size:|3.0 GB| + +## References + +https://huggingface.co/mlenjoyneer/rut5_large_sum_gazeta + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_ru.md b/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_ru.md new file mode 100644 index 00000000000000..2027c5ef3b6b52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-rut5_large_sum_gazeta_ru.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Russian rut5_large_sum_gazeta T5Transformer from mlenjoyneer +author: John Snow Labs +name: rut5_large_sum_gazeta +date: 2024-08-19 +tags: [ru, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_large_sum_gazeta` is a Russian model originally trained by mlenjoyneer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_large_sum_gazeta_ru_5.4.2_3.0_1724060009472.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_large_sum_gazeta_ru_5.4.2_3.0_1724060009472.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_large_sum_gazeta","ru") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_large_sum_gazeta", "ru") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_large_sum_gazeta| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ru| +|Size:|3.0 GB| + +## References + +https://huggingface.co/mlenjoyneer/rut5_large_sum_gazeta \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_en.md b/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_en.md new file mode 100644 index 00000000000000..56bf7abd810313 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sanskrit_tonga_tonga_islands_english T5Transformer from ubermenchh +author: John Snow Labs +name: sanskrit_tonga_tonga_islands_english +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sanskrit_tonga_tonga_islands_english` is a English model originally trained by ubermenchh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sanskrit_tonga_tonga_islands_english_en_5.4.2_3.0_1724026019969.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sanskrit_tonga_tonga_islands_english_en_5.4.2_3.0_1724026019969.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sanskrit_tonga_tonga_islands_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sanskrit_tonga_tonga_islands_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sanskrit_tonga_tonga_islands_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.6 MB| + +## References + +https://huggingface.co/ubermenchh/sanskrit-to-english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_pipeline_en.md new file mode 100644 index 00000000000000..83d490631b0520 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sanskrit_tonga_tonga_islands_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sanskrit_tonga_tonga_islands_english_pipeline pipeline T5Transformer from ubermenchh +author: John Snow Labs +name: sanskrit_tonga_tonga_islands_english_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sanskrit_tonga_tonga_islands_english_pipeline` is a English model originally trained by ubermenchh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sanskrit_tonga_tonga_islands_english_pipeline_en_5.4.2_3.0_1724026037805.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sanskrit_tonga_tonga_islands_english_pipeline_en_5.4.2_3.0_1724026037805.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sanskrit_tonga_tonga_islands_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sanskrit_tonga_tonga_islands_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sanskrit_tonga_tonga_islands_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.6 MB| + +## References + +https://huggingface.co/ubermenchh/sanskrit-to-english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_en.md b/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_en.md new file mode 100644 index 00000000000000..02eec1cb62693e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English semeval2023_clickbait_flan_t5_large_seed45 T5Transformer from tohokunlp-semeval2023-clickbait +author: John Snow Labs +name: semeval2023_clickbait_flan_t5_large_seed45 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`semeval2023_clickbait_flan_t5_large_seed45` is a English model originally trained by tohokunlp-semeval2023-clickbait. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/semeval2023_clickbait_flan_t5_large_seed45_en_5.4.2_3.0_1724029915383.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/semeval2023_clickbait_flan_t5_large_seed45_en_5.4.2_3.0_1724029915383.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("semeval2023_clickbait_flan_t5_large_seed45","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("semeval2023_clickbait_flan_t5_large_seed45", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|semeval2023_clickbait_flan_t5_large_seed45| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/tohokunlp-semeval2023-clickbait/semeval2023-clickbait-flan-t5-large-seed45 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_pipeline_en.md new file mode 100644 index 00000000000000..986c161a951379 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-semeval2023_clickbait_flan_t5_large_seed45_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English semeval2023_clickbait_flan_t5_large_seed45_pipeline pipeline T5Transformer from tohokunlp-semeval2023-clickbait +author: John Snow Labs +name: semeval2023_clickbait_flan_t5_large_seed45_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`semeval2023_clickbait_flan_t5_large_seed45_pipeline` is a English model originally trained by tohokunlp-semeval2023-clickbait. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/semeval2023_clickbait_flan_t5_large_seed45_pipeline_en_5.4.2_3.0_1724030049231.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/semeval2023_clickbait_flan_t5_large_seed45_pipeline_en_5.4.2_3.0_1724030049231.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("semeval2023_clickbait_flan_t5_large_seed45_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("semeval2023_clickbait_flan_t5_large_seed45_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|semeval2023_clickbait_flan_t5_large_seed45_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/tohokunlp-semeval2023-clickbait/semeval2023-clickbait-flan-t5-large-seed45 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_en.md b/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_en.md new file mode 100644 index 00000000000000..ab2f223f0ea96e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2 T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1724074786399.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_en_5.4.2_3.0_1724074786399.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|317.1 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline_en.md new file mode 100644 index 00000000000000..a6343bb168da5a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1724074804440.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline_en_5.4.2_3.0_1724074804440.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|317.1 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_en.md b/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_en.md new file mode 100644 index 00000000000000..085699131081d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English skill_extracting_t5_epoch2 T5Transformer from ManulaPankaja +author: John Snow Labs +name: skill_extracting_t5_epoch2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`skill_extracting_t5_epoch2` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/skill_extracting_t5_epoch2_en_5.4.2_3.0_1724105131160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/skill_extracting_t5_epoch2_en_5.4.2_3.0_1724105131160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("skill_extracting_t5_epoch2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("skill_extracting_t5_epoch2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|skill_extracting_t5_epoch2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|983.4 MB| + +## References + +https://huggingface.co/ManulaPankaja/skill_extracting_t5_epoch2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_pipeline_en.md new file mode 100644 index 00000000000000..927c340c85fc4d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-skill_extracting_t5_epoch2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English skill_extracting_t5_epoch2_pipeline pipeline T5Transformer from ManulaPankaja +author: John Snow Labs +name: skill_extracting_t5_epoch2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`skill_extracting_t5_epoch2_pipeline` is a English model originally trained by ManulaPankaja. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/skill_extracting_t5_epoch2_pipeline_en_5.4.2_3.0_1724105187782.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/skill_extracting_t5_epoch2_pipeline_en_5.4.2_3.0_1724105187782.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("skill_extracting_t5_epoch2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("skill_extracting_t5_epoch2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|skill_extracting_t5_epoch2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|983.4 MB| + +## References + +https://huggingface.co/ManulaPankaja/skill_extracting_t5_epoch2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_en.md b/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_en.md new file mode 100644 index 00000000000000..c399a210ff6db3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sled_lex_ukabs T5Transformer from CJWeiss +author: John Snow Labs +name: sled_lex_ukabs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sled_lex_ukabs` is a English model originally trained by CJWeiss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sled_lex_ukabs_en_5.4.2_3.0_1724091555323.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sled_lex_ukabs_en_5.4.2_3.0_1724091555323.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sled_lex_ukabs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sled_lex_ukabs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sled_lex_ukabs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/CJWeiss/sled_lex_ukabs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_pipeline_en.md new file mode 100644 index 00000000000000..c371a4beea0a94 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sled_lex_ukabs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sled_lex_ukabs_pipeline pipeline T5Transformer from CJWeiss +author: John Snow Labs +name: sled_lex_ukabs_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sled_lex_ukabs_pipeline` is a English model originally trained by CJWeiss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sled_lex_ukabs_pipeline_en_5.4.2_3.0_1724091601737.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sled_lex_ukabs_pipeline_en_5.4.2_3.0_1724091601737.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sled_lex_ukabs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sled_lex_ukabs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sled_lex_ukabs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/CJWeiss/sled_lex_ukabs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_en.md new file mode 100644 index 00000000000000..1b17ec654d56c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English small_t5_esnli_limited_efigsnli_e10_alpha_0_1 T5Transformer from harish +author: John Snow Labs +name: small_t5_esnli_limited_efigsnli_e10_alpha_0_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`small_t5_esnli_limited_efigsnli_e10_alpha_0_1` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/small_t5_esnli_limited_efigsnli_e10_alpha_0_1_en_5.4.2_3.0_1724095770821.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/small_t5_esnli_limited_efigsnli_e10_alpha_0_1_en_5.4.2_3.0_1724095770821.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("small_t5_esnli_limited_efigsnli_e10_alpha_0_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("small_t5_esnli_limited_efigsnli_e10_alpha_0_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|small_t5_esnli_limited_efigsnli_e10_alpha_0_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.5 MB| + +## References + +https://huggingface.co/harish/SMALL-t5-eSNLI-limited-eFigSNLI-e10-alpha-0-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline_en.md new file mode 100644 index 00000000000000..5455f17eb3a08d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline pipeline T5Transformer from harish +author: John Snow Labs +name: small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline` is a English model originally trained by harish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline_en_5.4.2_3.0_1724095795125.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline_en_5.4.2_3.0_1724095795125.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|small_t5_esnli_limited_efigsnli_e10_alpha_0_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.5 MB| + +## References + +https://huggingface.co/harish/SMALL-t5-eSNLI-limited-eFigSNLI-e10-alpha-0-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_en.md new file mode 100644 index 00000000000000..e1c900b9e46dba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English snli_t5_base_seed_2 T5Transformer from utahnlp +author: John Snow Labs +name: snli_t5_base_seed_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`snli_t5_base_seed_2` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/snli_t5_base_seed_2_en_5.4.2_3.0_1724088179930.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/snli_t5_base_seed_2_en_5.4.2_3.0_1724088179930.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("snli_t5_base_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("snli_t5_base_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|snli_t5_base_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|971.5 MB| + +## References + +https://huggingface.co/utahnlp/snli_t5-base_seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..55747abb18dded --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_base_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English snli_t5_base_seed_2_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: snli_t5_base_seed_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`snli_t5_base_seed_2_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/snli_t5_base_seed_2_pipeline_en_5.4.2_3.0_1724088240040.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/snli_t5_base_seed_2_pipeline_en_5.4.2_3.0_1724088240040.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("snli_t5_base_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("snli_t5_base_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|snli_t5_base_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|971.5 MB| + +## References + +https://huggingface.co/utahnlp/snli_t5-base_seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_en.md new file mode 100644 index 00000000000000..bf36e1ac98a01f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English snli_t5_small_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: snli_t5_small_seed_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`snli_t5_small_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/snli_t5_small_seed_1_en_5.4.2_3.0_1724060086868.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/snli_t5_small_seed_1_en_5.4.2_3.0_1724060086868.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("snli_t5_small_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("snli_t5_small_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|snli_t5_small_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|329.4 MB| + +## References + +https://huggingface.co/utahnlp/snli_t5-small_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..1e5d0e8799fc09 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-snli_t5_small_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English snli_t5_small_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: snli_t5_small_seed_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`snli_t5_small_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/snli_t5_small_seed_1_pipeline_en_5.4.2_3.0_1724060107696.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/snli_t5_small_seed_1_pipeline_en_5.4.2_3.0_1724060107696.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("snli_t5_small_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("snli_t5_small_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|snli_t5_small_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|329.4 MB| + +## References + +https://huggingface.co/utahnlp/snli_t5-small_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_en.md b/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_en.md new file mode 100644 index 00000000000000..f24194086ec324 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English spanish_uc_modelo_npl_e3 T5Transformer from JuanForeroNeme +author: John Snow Labs +name: spanish_uc_modelo_npl_e3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`spanish_uc_modelo_npl_e3` is a English model originally trained by JuanForeroNeme. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/spanish_uc_modelo_npl_e3_en_5.4.2_3.0_1724111820656.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/spanish_uc_modelo_npl_e3_en_5.4.2_3.0_1724111820656.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("spanish_uc_modelo_npl_e3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("spanish_uc_modelo_npl_e3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|spanish_uc_modelo_npl_e3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/JuanForeroNeme/ES_UC_MODELO_NPL_E3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_pipeline_en.md new file mode 100644 index 00000000000000..1cdb61ba483159 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-spanish_uc_modelo_npl_e3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English spanish_uc_modelo_npl_e3_pipeline pipeline T5Transformer from JuanForeroNeme +author: John Snow Labs +name: spanish_uc_modelo_npl_e3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`spanish_uc_modelo_npl_e3_pipeline` is a English model originally trained by JuanForeroNeme. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/spanish_uc_modelo_npl_e3_pipeline_en_5.4.2_3.0_1724111870319.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/spanish_uc_modelo_npl_e3_pipeline_en_5.4.2_3.0_1724111870319.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("spanish_uc_modelo_npl_e3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("spanish_uc_modelo_npl_e3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|spanish_uc_modelo_npl_e3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/JuanForeroNeme/ES_UC_MODELO_NPL_E3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_en.md b/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_en.md new file mode 100644 index 00000000000000..f44b97cae479b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English speech_chatgpt_long_t5_base_encodec2instruction_prompttts T5Transformer from kuanhuggingface +author: John Snow Labs +name: speech_chatgpt_long_t5_base_encodec2instruction_prompttts +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`speech_chatgpt_long_t5_base_encodec2instruction_prompttts` is a English model originally trained by kuanhuggingface. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/speech_chatgpt_long_t5_base_encodec2instruction_prompttts_en_5.4.2_3.0_1724074182055.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/speech_chatgpt_long_t5_base_encodec2instruction_prompttts_en_5.4.2_3.0_1724074182055.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("speech_chatgpt_long_t5_base_encodec2instruction_prompttts","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("speech_chatgpt_long_t5_base_encodec2instruction_prompttts", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|speech_chatgpt_long_t5_base_encodec2instruction_prompttts| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/kuanhuggingface/speech-chatgpt-long-t5-base-encodec2instruction-promptTTS \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline_en.md new file mode 100644 index 00000000000000..2ba76a1de04798 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline pipeline T5Transformer from kuanhuggingface +author: John Snow Labs +name: speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline` is a English model originally trained by kuanhuggingface. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline_en_5.4.2_3.0_1724074236871.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline_en_5.4.2_3.0_1724074236871.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|speech_chatgpt_long_t5_base_encodec2instruction_prompttts_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/kuanhuggingface/speech-chatgpt-long-t5-base-encodec2instruction-promptTTS + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_en.md b/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_en.md new file mode 100644 index 00000000000000..716fdf3ae1d4c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English spell_corrector_mt5_01012024 T5Transformer from Buseak +author: John Snow Labs +name: spell_corrector_mt5_01012024 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`spell_corrector_mt5_01012024` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/spell_corrector_mt5_01012024_en_5.4.2_3.0_1724045598603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/spell_corrector_mt5_01012024_en_5.4.2_3.0_1724045598603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("spell_corrector_mt5_01012024","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("spell_corrector_mt5_01012024", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|spell_corrector_mt5_01012024| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/spell_corrector_mt5_01012024 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_pipeline_en.md new file mode 100644 index 00000000000000..bdd2419f1f6b0c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-spell_corrector_mt5_01012024_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English spell_corrector_mt5_01012024_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: spell_corrector_mt5_01012024_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`spell_corrector_mt5_01012024_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/spell_corrector_mt5_01012024_pipeline_en_5.4.2_3.0_1724045754382.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/spell_corrector_mt5_01012024_pipeline_en_5.4.2_3.0_1724045754382.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("spell_corrector_mt5_01012024_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("spell_corrector_mt5_01012024_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|spell_corrector_mt5_01012024_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/spell_corrector_mt5_01012024 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_en.md b/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_en.md new file mode 100644 index 00000000000000..fa4886ef54dc14 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sql_kaggle T5Transformer from AayushShah +author: John Snow Labs +name: sql_kaggle +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sql_kaggle` is a English model originally trained by AayushShah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sql_kaggle_en_5.4.2_3.0_1724026668729.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sql_kaggle_en_5.4.2_3.0_1724026668729.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sql_kaggle","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sql_kaggle", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sql_kaggle| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/AayushShah/SQL_Kaggle \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_pipeline_en.md new file mode 100644 index 00000000000000..c043282b547fbe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sql_kaggle_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sql_kaggle_pipeline pipeline T5Transformer from AayushShah +author: John Snow Labs +name: sql_kaggle_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sql_kaggle_pipeline` is a English model originally trained by AayushShah. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sql_kaggle_pipeline_en_5.4.2_3.0_1724026715154.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sql_kaggle_pipeline_en_5.4.2_3.0_1724026715154.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sql_kaggle_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sql_kaggle_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sql_kaggle_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/AayushShah/SQL_Kaggle + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_en.md b/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_en.md new file mode 100644 index 00000000000000..b3c11223ab93a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English squad_bengali_qgen_banglat5 T5Transformer from jannatul17 +author: John Snow Labs +name: squad_bengali_qgen_banglat5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`squad_bengali_qgen_banglat5` is a English model originally trained by jannatul17. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/squad_bengali_qgen_banglat5_en_5.4.2_3.0_1724059873945.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/squad_bengali_qgen_banglat5_en_5.4.2_3.0_1724059873945.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("squad_bengali_qgen_banglat5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("squad_bengali_qgen_banglat5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|squad_bengali_qgen_banglat5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|995.9 MB| + +## References + +https://huggingface.co/jannatul17/squad-bn-qgen-banglat5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_pipeline_en.md new file mode 100644 index 00000000000000..c42b8610de607e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-squad_bengali_qgen_banglat5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English squad_bengali_qgen_banglat5_pipeline pipeline T5Transformer from jannatul17 +author: John Snow Labs +name: squad_bengali_qgen_banglat5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`squad_bengali_qgen_banglat5_pipeline` is a English model originally trained by jannatul17. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/squad_bengali_qgen_banglat5_pipeline_en_5.4.2_3.0_1724059924023.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/squad_bengali_qgen_banglat5_pipeline_en_5.4.2_3.0_1724059924023.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("squad_bengali_qgen_banglat5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("squad_bengali_qgen_banglat5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|squad_bengali_qgen_banglat5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|995.9 MB| + +## References + +https://huggingface.co/jannatul17/squad-bn-qgen-banglat5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-squad_en.md b/docs/_posts/ahmedlone127/2024-08-19-squad_en.md new file mode 100644 index 00000000000000..0e15ed6253489e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-squad_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English squad T5Transformer from Rhasan97 +author: John Snow Labs +name: squad +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`squad` is a English model originally trained by Rhasan97. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/squad_en_5.4.2_3.0_1724086357634.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/squad_en_5.4.2_3.0_1724086357634.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("squad","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("squad", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|squad| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/Rhasan97/squad \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-squad_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-squad_pipeline_en.md new file mode 100644 index 00000000000000..ee3322031157a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-squad_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English squad_pipeline pipeline T5Transformer from Rhasan97 +author: John Snow Labs +name: squad_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`squad_pipeline` is a English model originally trained by Rhasan97. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/squad_pipeline_en_5.4.2_3.0_1724086375160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/squad_pipeline_en_5.4.2_3.0_1724086375160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("squad_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("squad_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|squad_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/Rhasan97/squad + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_en.md b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_en.md new file mode 100644 index 00000000000000..3e4e90060dd9df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sum_model_lr1e_3_20_30epoch T5Transformer from weny22 +author: John Snow Labs +name: sum_model_lr1e_3_20_30epoch +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_model_lr1e_3_20_30epoch` is a English model originally trained by weny22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_model_lr1e_3_20_30epoch_en_5.4.2_3.0_1724094016580.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_model_lr1e_3_20_30epoch_en_5.4.2_3.0_1724094016580.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sum_model_lr1e_3_20_30epoch","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sum_model_lr1e_3_20_30epoch", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_model_lr1e_3_20_30epoch| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|420.4 MB| + +## References + +https://huggingface.co/weny22/sum_model_lr1e_3_20_30epoch \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_pipeline_en.md new file mode 100644 index 00000000000000..6ba757ba3a70cb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr1e_3_20_30epoch_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sum_model_lr1e_3_20_30epoch_pipeline pipeline T5Transformer from weny22 +author: John Snow Labs +name: sum_model_lr1e_3_20_30epoch_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_model_lr1e_3_20_30epoch_pipeline` is a English model originally trained by weny22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_model_lr1e_3_20_30epoch_pipeline_en_5.4.2_3.0_1724094036817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_model_lr1e_3_20_30epoch_pipeline_en_5.4.2_3.0_1724094036817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sum_model_lr1e_3_20_30epoch_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sum_model_lr1e_3_20_30epoch_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_model_lr1e_3_20_30epoch_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|420.4 MB| + +## References + +https://huggingface.co/weny22/sum_model_lr1e_3_20_30epoch + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_en.md b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_en.md new file mode 100644 index 00000000000000..394c7731364d5c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English sum_model_lr2e_4_20epoch T5Transformer from weny22 +author: John Snow Labs +name: sum_model_lr2e_4_20epoch +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_model_lr2e_4_20epoch` is a English model originally trained by weny22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_model_lr2e_4_20epoch_en_5.4.2_3.0_1724111586773.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_model_lr2e_4_20epoch_en_5.4.2_3.0_1724111586773.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("sum_model_lr2e_4_20epoch","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("sum_model_lr2e_4_20epoch", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_model_lr2e_4_20epoch| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|420.4 MB| + +## References + +https://huggingface.co/weny22/sum_model_lr2e_4_20epoch \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_pipeline_en.md new file mode 100644 index 00000000000000..446c70dbf97226 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-sum_model_lr2e_4_20epoch_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English sum_model_lr2e_4_20epoch_pipeline pipeline T5Transformer from weny22 +author: John Snow Labs +name: sum_model_lr2e_4_20epoch_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`sum_model_lr2e_4_20epoch_pipeline` is a English model originally trained by weny22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/sum_model_lr2e_4_20epoch_pipeline_en_5.4.2_3.0_1724111607215.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/sum_model_lr2e_4_20epoch_pipeline_en_5.4.2_3.0_1724111607215.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("sum_model_lr2e_4_20epoch_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("sum_model_lr2e_4_20epoch_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|sum_model_lr2e_4_20epoch_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|420.4 MB| + +## References + +https://huggingface.co/weny22/sum_model_lr2e_4_20epoch + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_en.md new file mode 100644 index 00000000000000..10cae2f191fd8d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarization_model_test_full T5Transformer from jnelen +author: John Snow Labs +name: summarization_model_test_full +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_model_test_full` is a English model originally trained by jnelen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_model_test_full_en_5.4.2_3.0_1724026819479.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_model_test_full_en_5.4.2_3.0_1724026819479.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarization_model_test_full","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarization_model_test_full", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_model_test_full| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/jnelen/summarization_model_test_full \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_pipeline_en.md new file mode 100644 index 00000000000000..e3a348ad5728ad --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarization_model_test_full_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarization_model_test_full_pipeline pipeline T5Transformer from jnelen +author: John Snow Labs +name: summarization_model_test_full_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_model_test_full_pipeline` is a English model originally trained by jnelen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_model_test_full_pipeline_en_5.4.2_3.0_1724026835708.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_model_test_full_pipeline_en_5.4.2_3.0_1724026835708.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarization_model_test_full_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarization_model_test_full_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_model_test_full_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/jnelen/summarization_model_test_full + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_en.md new file mode 100644 index 00000000000000..b46d0a79cb5fba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarization_tfm_claudiasoria_v4 T5Transformer from clxudiajazmin +author: John Snow Labs +name: summarization_tfm_claudiasoria_v4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_tfm_claudiasoria_v4` is a English model originally trained by clxudiajazmin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_tfm_claudiasoria_v4_en_5.4.2_3.0_1724092072464.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_tfm_claudiasoria_v4_en_5.4.2_3.0_1724092072464.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarization_tfm_claudiasoria_v4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarization_tfm_claudiasoria_v4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_tfm_claudiasoria_v4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.2 MB| + +## References + +https://huggingface.co/clxudiajazmin/summarization-TFM-ClaudiaSoria_v4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_pipeline_en.md new file mode 100644 index 00000000000000..7ca82e6e92e62b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarization_tfm_claudiasoria_v4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarization_tfm_claudiasoria_v4_pipeline pipeline T5Transformer from clxudiajazmin +author: John Snow Labs +name: summarization_tfm_claudiasoria_v4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarization_tfm_claudiasoria_v4_pipeline` is a English model originally trained by clxudiajazmin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarization_tfm_claudiasoria_v4_pipeline_en_5.4.2_3.0_1724092090540.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarization_tfm_claudiasoria_v4_pipeline_en_5.4.2_3.0_1724092090540.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarization_tfm_claudiasoria_v4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarization_tfm_claudiasoria_v4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarization_tfm_claudiasoria_v4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.2 MB| + +## References + +https://huggingface.co/clxudiajazmin/summarization-TFM-ClaudiaSoria_v4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_en.md new file mode 100644 index 00000000000000..79d18ad0f50efd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarizer_google_long_t5_tglobal_base_keywords_faceted T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_tglobal_base_keywords_faceted +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_tglobal_base_keywords_faceted` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_keywords_faceted_en_5.4.2_3.0_1724081781095.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_keywords_faceted_en_5.4.2_3.0_1724081781095.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarizer_google_long_t5_tglobal_base_keywords_faceted","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarizer_google_long_t5_tglobal_base_keywords_faceted", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_tglobal_base_keywords_faceted| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-tglobal-base_keywords_faceted \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline_en.md new file mode 100644 index 00000000000000..7111024622425e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline pipeline T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline_en_5.4.2_3.0_1724081827905.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline_en_5.4.2_3.0_1724081827905.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_tglobal_base_keywords_faceted_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-tglobal-base_keywords_faceted + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_en.md new file mode 100644 index 00000000000000..5aadec77b0d131 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_en_5.4.2_3.0_1724093198992.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_en_5.4.2_3.0_1724093198992.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-tglobal-base_most_frequent_unfaceted \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline_en.md new file mode 100644 index 00000000000000..c7c629b6a4eb31 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline pipeline T5Transformer from acmc +author: John Snow Labs +name: summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline` is a English model originally trained by acmc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline_en_5.4.2_3.0_1724093250056.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline_en_5.4.2_3.0_1724093250056.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summarizer_google_long_t5_tglobal_base_most_frequent_unfaceted_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acmc/summarizer_google_long-t5-tglobal-base_most_frequent_unfaceted + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_en.md b/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_en.md new file mode 100644 index 00000000000000..60c6513a644bd6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summary_cz_eurlex T5Transformer from Anopheles +author: John Snow Labs +name: summary_cz_eurlex +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summary_cz_eurlex` is a English model originally trained by Anopheles. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summary_cz_eurlex_en_5.4.2_3.0_1724075856746.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summary_cz_eurlex_en_5.4.2_3.0_1724075856746.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summary_cz_eurlex","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summary_cz_eurlex", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summary_cz_eurlex| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|276.9 MB| + +## References + +https://huggingface.co/Anopheles/summary_cz_eurlex \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_pipeline_en.md new file mode 100644 index 00000000000000..5ee060fb05507c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summary_cz_eurlex_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summary_cz_eurlex_pipeline pipeline T5Transformer from Anopheles +author: John Snow Labs +name: summary_cz_eurlex_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summary_cz_eurlex_pipeline` is a English model originally trained by Anopheles. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summary_cz_eurlex_pipeline_en_5.4.2_3.0_1724075887401.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summary_cz_eurlex_pipeline_en_5.4.2_3.0_1724075887401.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summary_cz_eurlex_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summary_cz_eurlex_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summary_cz_eurlex_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|276.9 MB| + +## References + +https://huggingface.co/Anopheles/summary_cz_eurlex + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_en.md b/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_en.md new file mode 100644 index 00000000000000..c1a931d5eb2f73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English summary_of_news_articles_nepal_bhasa T5Transformer from therapara +author: John Snow Labs +name: summary_of_news_articles_nepal_bhasa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summary_of_news_articles_nepal_bhasa` is a English model originally trained by therapara. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summary_of_news_articles_nepal_bhasa_en_5.4.2_3.0_1724046724926.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summary_of_news_articles_nepal_bhasa_en_5.4.2_3.0_1724046724926.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("summary_of_news_articles_nepal_bhasa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("summary_of_news_articles_nepal_bhasa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summary_of_news_articles_nepal_bhasa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.7 MB| + +## References + +https://huggingface.co/therapara/summary-of-news-articles_new \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_pipeline_en.md new file mode 100644 index 00000000000000..8f323516958d19 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-summary_of_news_articles_nepal_bhasa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English summary_of_news_articles_nepal_bhasa_pipeline pipeline T5Transformer from therapara +author: John Snow Labs +name: summary_of_news_articles_nepal_bhasa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`summary_of_news_articles_nepal_bhasa_pipeline` is a English model originally trained by therapara. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/summary_of_news_articles_nepal_bhasa_pipeline_en_5.4.2_3.0_1724046742504.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/summary_of_news_articles_nepal_bhasa_pipeline_en_5.4.2_3.0_1724046742504.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("summary_of_news_articles_nepal_bhasa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("summary_of_news_articles_nepal_bhasa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|summary_of_news_articles_nepal_bhasa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.7 MB| + +## References + +https://huggingface.co/therapara/summary-of-news-articles_new + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_en.md b/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_en.md new file mode 100644 index 00000000000000..2aca2c41c04887 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English superglue_copa T5Transformer from ShengdingHu +author: John Snow Labs +name: superglue_copa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`superglue_copa` is a English model originally trained by ShengdingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/superglue_copa_en_5.4.2_3.0_1724033289230.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/superglue_copa_en_5.4.2_3.0_1724033289230.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("superglue_copa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("superglue_copa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|superglue_copa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ShengdingHu/superglue-copa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_pipeline_en.md new file mode 100644 index 00000000000000..97a4b0976d63fa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-superglue_copa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English superglue_copa_pipeline pipeline T5Transformer from ShengdingHu +author: John Snow Labs +name: superglue_copa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`superglue_copa_pipeline` is a English model originally trained by ShengdingHu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/superglue_copa_pipeline_en_5.4.2_3.0_1724033336799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/superglue_copa_pipeline_en_5.4.2_3.0_1724033336799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("superglue_copa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("superglue_copa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|superglue_copa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ShengdingHu/superglue-copa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5150_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5150_en.md new file mode 100644 index 00000000000000..a1ec82d4843180 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5150_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5150 T5Transformer from Saddammm +author: John Snow Labs +name: t5150 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5150` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5150_en_5.4.2_3.0_1724090663856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5150_en_5.4.2_3.0_1724090663856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5150","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5150", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5150| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/T5150 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5150_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5150_pipeline_en.md new file mode 100644 index 00000000000000..28c44e99d02000 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5150_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5150_pipeline pipeline T5Transformer from Saddammm +author: John Snow Labs +name: t5150_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5150_pipeline` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5150_pipeline_en_5.4.2_3.0_1724090712077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5150_pipeline_en_5.4.2_3.0_1724090712077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5150_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5150_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5150_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/T5150 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_en.md new file mode 100644 index 00000000000000..2ba7d615946258 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2013_0 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2013_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2013_0` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_0_en_5.4.2_3.0_1724044104494.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_0_en_5.4.2_3.0_1724044104494.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2013_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2013_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2013_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2013-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_pipeline_en.md new file mode 100644 index 00000000000000..c1855e0625c085 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2013_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2013_0_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2013_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2013_0_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_0_pipeline_en_5.4.2_3.0_1724044121439.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2013_0_pipeline_en_5.4.2_3.0_1724044121439.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2013_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2013_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2013_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2013-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_en.md new file mode 100644 index 00000000000000..f9ecd56510332f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2015_5 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2015_5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2015_5` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_5_en_5.4.2_3.0_1724100391725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_5_en_5.4.2_3.0_1724100391725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2015_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2015_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2015_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2015-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_pipeline_en.md new file mode 100644 index 00000000000000..41a795d3f02416 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2015_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2015_5_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2015_5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2015_5_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_5_pipeline_en_5.4.2_3.0_1724100409197.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_5_pipeline_en_5.4.2_3.0_1724100409197.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2015_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2015_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2015_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2015-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_en.md new file mode 100644 index 00000000000000..e89b8c08e1c851 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_9 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_9 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_9` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_9_en_5.4.2_3.0_1724038572278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_9_en_5.4.2_3.0_1724038572278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_9","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_9", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_9| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-9 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_pipeline_en.md new file mode 100644 index 00000000000000..1bac0d847095ff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2018_9_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_9_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_9_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_9_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_9_pipeline_en_5.4.2_3.0_1724038588511.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_9_pipeline_en_5.4.2_3.0_1724038588511.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2018_9_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2018_9_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_9_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-9 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_en.md new file mode 100644 index 00000000000000..f5b63d5ac00c23 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2019_6 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2019_6 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2019_6` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2019_6_en_5.4.2_3.0_1724027342609.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2019_6_en_5.4.2_3.0_1724027342609.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2019_6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2019_6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2019_6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2019-6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_pipeline_en.md new file mode 100644 index 00000000000000..e18d12c3f88ebb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2019_6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2019_6_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2019_6_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2019_6_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2019_6_pipeline_en_5.4.2_3.0_1724027359143.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2019_6_pipeline_en_5.4.2_3.0_1724027359143.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2019_6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2019_6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2019_6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2019-6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_en.md new file mode 100644 index 00000000000000..46966806a239b9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2020_10 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2020_10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2020_10` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2020_10_en_5.4.2_3.0_1724100790259.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2020_10_en_5.4.2_3.0_1724100790259.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2020_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2020_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2020_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2020-10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_pipeline_en.md new file mode 100644 index 00000000000000..cfae27766acc74 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_lm_wmt_2020_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2020_10_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2020_10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2020_10_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2020_10_pipeline_en_5.4.2_3.0_1724100807984.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2020_10_pipeline_en_5.4.2_3.0_1724100807984.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2020_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2020_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2020_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.0 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2020-10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_en.md new file mode 100644 index 00000000000000..06cdc0f822a600 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_news_cls_2012 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_news_cls_2012 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_news_cls_2012` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_2012_en_5.4.2_3.0_1724100585234.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_2012_en_5.4.2_3.0_1724100585234.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_news_cls_2012","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_news_cls_2012", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_news_cls_2012| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-news_cls-2012 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_pipeline_en.md new file mode 100644 index 00000000000000..61605a6c0ff560 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_2012_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_news_cls_2012_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_news_cls_2012_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_news_cls_2012_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_2012_pipeline_en_5.4.2_3.0_1724100605187.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_2012_pipeline_en_5.4.2_3.0_1724100605187.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_news_cls_2012_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_news_cls_2012_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_news_cls_2012_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-news_cls-2012 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_en.md new file mode 100644 index 00000000000000..06064b311ee2be --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_news_cls_combined_years T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_news_cls_combined_years +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_news_cls_combined_years` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_combined_years_en_5.4.2_3.0_1724094539819.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_combined_years_en_5.4.2_3.0_1724094539819.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_news_cls_combined_years","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_news_cls_combined_years", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_news_cls_combined_years| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-news_cls-combined_years \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_pipeline_en.md new file mode 100644 index 00000000000000..84a9dd54b1a13b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_news_cls_combined_years_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_news_cls_combined_years_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_news_cls_combined_years_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_news_cls_combined_years_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_combined_years_pipeline_en_5.4.2_3.0_1724094559980.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_news_cls_combined_years_pipeline_en_5.4.2_3.0_1724094559980.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_news_cls_combined_years_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_news_cls_combined_years_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_news_cls_combined_years_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-news_cls-combined_years + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_en.md new file mode 100644 index 00000000000000..732d841bb16136 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_2 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_2` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_2_en_5.4.2_3.0_1724100353356.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_2_en_5.4.2_3.0_1724100353356.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|297.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_pipeline_en.md new file mode 100644 index 00000000000000..2851c72db2fa92 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_2_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_2_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_2_pipeline_en_5.4.2_3.0_1724100383107.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_2_pipeline_en_5.4.2_3.0_1724100383107.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2016_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2016_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|297.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_en.md new file mode 100644 index 00000000000000..14c95397b9d2eb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_3 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_3` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_3_en_5.4.2_3.0_1724091512858.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_3_en_5.4.2_3.0_1724091512858.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|301.9 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_pipeline_en.md new file mode 100644 index 00000000000000..1152d81b0de5dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_3_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_3_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_3_pipeline_en_5.4.2_3.0_1724091540512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_3_pipeline_en_5.4.2_3.0_1724091540512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2016_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2016_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|301.9 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_en.md new file mode 100644 index 00000000000000..686583872a0c71 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_5 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_5` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_5_en_5.4.2_3.0_1724086835346.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_5_en_5.4.2_3.0_1724086835346.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2016_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|303.4 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_pipeline_en.md new file mode 100644 index 00000000000000..b1959845f1c90a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2016_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2016_5_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2016_5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2016_5_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_5_pipeline_en_5.4.2_3.0_1724086867832.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2016_5_pipeline_en_5.4.2_3.0_1724086867832.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2016_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2016_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2016_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|303.4 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2016-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_en.md new file mode 100644 index 00000000000000..370b3fb00d2a42 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_1 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_1` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_1_en_5.4.2_3.0_1724036390375.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_1_en_5.4.2_3.0_1724036390375.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|297.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_pipeline_en.md new file mode 100644 index 00000000000000..6aab82b0cb5dca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2017_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_1_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_1_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_1_pipeline_en_5.4.2_3.0_1724036418529.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_1_pipeline_en_5.4.2_3.0_1724036418529.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2017_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2017_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|297.8 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_en.md new file mode 100644 index 00000000000000..171bb0126eeca9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_1 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_1` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_1_en_5.4.2_3.0_1724090591405.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_1_en_5.4.2_3.0_1724090591405.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|299.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_pipeline_en.md new file mode 100644 index 00000000000000..92a83a041ed081 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_1_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_1_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_1_pipeline_en_5.4.2_3.0_1724090621181.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_1_pipeline_en_5.4.2_3.0_1724090621181.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2018_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2018_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|299.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_en.md new file mode 100644 index 00000000000000..ed4238e50e49aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_9 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_9 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_9` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_9_en_5.4.2_3.0_1724033181114.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_9_en_5.4.2_3.0_1724033181114.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_9","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_9", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_9| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|302.4 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-9 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_pipeline_en.md new file mode 100644 index 00000000000000..2d29b0e79e1df3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2018_9_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_9_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_9_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_9_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_9_pipeline_en_5.4.2_3.0_1724033208711.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_9_pipeline_en_5.4.2_3.0_1724033208711.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2018_9_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2018_9_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_9_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|302.4 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-9 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_en.md new file mode 100644 index 00000000000000..f3511f73442296 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2019_0 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2019_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2019_0` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_0_en_5.4.2_3.0_1724070846725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_0_en_5.4.2_3.0_1724070846725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2019_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2019_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2019_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|303.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2019-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_pipeline_en.md new file mode 100644 index 00000000000000..7ce3880527dc66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_2019_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2019_0_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2019_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2019_0_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_0_pipeline_en_5.4.2_3.0_1724070875166.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2019_0_pipeline_en_5.4.2_3.0_1724070875166.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2019_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2019_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2019_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|303.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2019-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_en.md new file mode 100644 index 00000000000000..6c5e4d92dbb565 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_combined_years_with_year_flag T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_combined_years_with_year_flag +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_combined_years_with_year_flag` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_combined_years_with_year_flag_en_5.4.2_3.0_1724032947244.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_combined_years_with_year_flag_en_5.4.2_3.0_1724032947244.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_combined_years_with_year_flag","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_combined_years_with_year_flag", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_combined_years_with_year_flag| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-combined_years_with_year_flag \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_pipeline_en.md new file mode 100644 index 00000000000000..d1ef294f6b0625 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_60m_poli_aff_combined_years_with_year_flag_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_combined_years_with_year_flag_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_combined_years_with_year_flag_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_combined_years_with_year_flag_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_combined_years_with_year_flag_pipeline_en_5.4.2_3.0_1724032967443.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_combined_years_with_year_flag_pipeline_en_5.4.2_3.0_1724032967443.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_combined_years_with_year_flag_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_combined_years_with_year_flag_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_combined_years_with_year_flag_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-combined_years_with_year_flag + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_en.md new file mode 100644 index 00000000000000..121f0f9027bccc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_argument_anlyser T5Transformer from addy88 +author: John Snow Labs +name: t5_argument_anlyser +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_argument_anlyser` is a English model originally trained by addy88. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_argument_anlyser_en_5.4.2_3.0_1724037162553.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_argument_anlyser_en_5.4.2_3.0_1724037162553.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_argument_anlyser","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_argument_anlyser", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_argument_anlyser| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/addy88/t5-argument-anlyser \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_pipeline_en.md new file mode 100644 index 00000000000000..fcd915de673230 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_argument_anlyser_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_argument_anlyser_pipeline pipeline T5Transformer from addy88 +author: John Snow Labs +name: t5_argument_anlyser_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_argument_anlyser_pipeline` is a English model originally trained by addy88. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_argument_anlyser_pipeline_en_5.4.2_3.0_1724037308103.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_argument_anlyser_pipeline_en_5.4.2_3.0_1724037308103.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_argument_anlyser_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_argument_anlyser_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_argument_anlyser_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/addy88/t5-argument-anlyser + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_en.md new file mode 100644 index 00000000000000..e0211a29fd9254 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_asqa_ob_irenepap T5Transformer from irenepap +author: John Snow Labs +name: t5_base_asqa_ob_irenepap +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_asqa_ob_irenepap` is a English model originally trained by irenepap. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_asqa_ob_irenepap_en_5.4.2_3.0_1724060612854.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_asqa_ob_irenepap_en_5.4.2_3.0_1724060612854.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_asqa_ob_irenepap","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_asqa_ob_irenepap", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_asqa_ob_irenepap| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|994.0 MB| + +## References + +https://huggingface.co/irenepap/t5-base-asqa-ob \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_pipeline_en.md new file mode 100644 index 00000000000000..7ed1fa7c9c1a94 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_asqa_ob_irenepap_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_asqa_ob_irenepap_pipeline pipeline T5Transformer from irenepap +author: John Snow Labs +name: t5_base_asqa_ob_irenepap_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_asqa_ob_irenepap_pipeline` is a English model originally trained by irenepap. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_asqa_ob_irenepap_pipeline_en_5.4.2_3.0_1724060662799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_asqa_ob_irenepap_pipeline_en_5.4.2_3.0_1724060662799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_asqa_ob_irenepap_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_asqa_ob_irenepap_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_asqa_ob_irenepap_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|994.0 MB| + +## References + +https://huggingface.co/irenepap/t5-base-asqa-ob + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_en.md new file mode 100644 index 00000000000000..669fe6ab90b28c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt4 T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_en_5.4.2_3.0_1724105692478.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_en_5.4.2_3.0_1724105692478.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_en.md new file mode 100644 index 00000000000000..0d5e82b00e6beb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt4_khanq_eduqg T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4_khanq_eduqg +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4_khanq_eduqg` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_eduqg_en_5.4.2_3.0_1724072806808.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_eduqg_en_5.4.2_3.0_1724072806808.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt4_khanq_eduqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt4_khanq_eduqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4_khanq_eduqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4-khanq-eduqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_pipeline_en.md new file mode 100644 index 00000000000000..a8691418d009c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_eduqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt4_khanq_eduqg_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4_khanq_eduqg_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4_khanq_eduqg_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_eduqg_pipeline_en_5.4.2_3.0_1724072856924.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_eduqg_pipeline_en_5.4.2_3.0_1724072856924.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt4_khanq_eduqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt4_khanq_eduqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4_khanq_eduqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4-khanq-eduqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_en.md new file mode 100644 index 00000000000000..b2ea8e419cb151 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt4_khanq T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4_khanq +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4_khanq` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_en_5.4.2_3.0_1724043460021.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_en_5.4.2_3.0_1724043460021.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt4_khanq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt4_khanq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4_khanq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4-khanq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_pipeline_en.md new file mode 100644 index 00000000000000..5a3b990f6d5248 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_khanq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt4_khanq_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4_khanq_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4_khanq_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_pipeline_en_5.4.2_3.0_1724043507623.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_khanq_pipeline_en_5.4.2_3.0_1724043507623.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt4_khanq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt4_khanq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4_khanq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4-khanq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_pipeline_en.md new file mode 100644 index 00000000000000..74805a567a081d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt4_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt4_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt4_pipeline_en_5.4.2_3.0_1724105739723.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt4_pipeline_en_5.4.2_3.0_1724105739723.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_en.md new file mode 100644 index 00000000000000..29195d3437c314 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt5_khanq T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt5_khanq +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt5_khanq` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt5_khanq_en_5.4.2_3.0_1724101952703.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt5_khanq_en_5.4.2_3.0_1724101952703.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt5_khanq","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt5_khanq", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt5_khanq| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt5-khanq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_pipeline_en.md new file mode 100644 index 00000000000000..c83eab490f003a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_bt5_khanq_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt5_khanq_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt5_khanq_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt5_khanq_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt5_khanq_pipeline_en_5.4.2_3.0_1724102001717.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt5_khanq_pipeline_en_5.4.2_3.0_1724102001717.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt5_khanq_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt5_khanq_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt5_khanq_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt5-khanq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_en.md new file mode 100644 index 00000000000000..cf8aa756569091 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_en_5.4.2_3.0_1724071363616.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_en_5.4.2_3.0_1724071363616.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline_en.md new file mode 100644 index 00000000000000..e612d2967012e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline pipeline T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline_en_5.4.2_3.0_1724071514302.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline_en_5.4.2_3.0_1724071514302.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_c2_mare_ar1_ex8_half_from_ft_dense_with_sort_noise_scaler + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_en.md new file mode 100644 index 00000000000000..67dfb337e75477 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_danish_multiwoz2_1_500 T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_danish_multiwoz2_1_500 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_danish_multiwoz2_1_500` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_danish_multiwoz2_1_500_en_5.4.2_3.0_1724060680670.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_danish_multiwoz2_1_500_en_5.4.2_3.0_1724060680670.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_danish_multiwoz2_1_500","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_danish_multiwoz2_1_500", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_danish_multiwoz2_1_500| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|939.2 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-da-multiwoz2.1_500 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_pipeline_en.md new file mode 100644 index 00000000000000..6458c7ce956a64 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_danish_multiwoz2_1_500_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_danish_multiwoz2_1_500_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_danish_multiwoz2_1_500_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_danish_multiwoz2_1_500_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_danish_multiwoz2_1_500_pipeline_en_5.4.2_3.0_1724060734122.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_danish_multiwoz2_1_500_pipeline_en_5.4.2_3.0_1724060734122.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_danish_multiwoz2_1_500_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_danish_multiwoz2_1_500_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_danish_multiwoz2_1_500_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|939.2 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-da-multiwoz2.1_500 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_en.md new file mode 100644 index 00000000000000..6ebf92f0601097 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_512_finetuned_squad_seed_0 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_512_finetuned_squad_seed_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_512_finetuned_squad_seed_0` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_512_finetuned_squad_seed_0_en_5.4.2_3.0_1724030814858.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_512_finetuned_squad_seed_0_en_5.4.2_3.0_1724030814858.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_512_finetuned_squad_seed_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_512_finetuned_squad_seed_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_512_finetuned_squad_seed_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|965.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-512-finetuned-squad-seed-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline_en.md new file mode 100644 index 00000000000000..29ee761395ea0c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline_en_5.4.2_3.0_1724030879139.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline_en_5.4.2_3.0_1724030879139.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_512_finetuned_squad_seed_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|965.1 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-512-finetuned-squad-seed-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_en.md new file mode 100644 index 00000000000000..952def80e912d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_bbc_furyhawk T5Transformer from furyhawk +author: John Snow Labs +name: t5_base_finetuned_bbc_furyhawk +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_bbc_furyhawk` is a English model originally trained by furyhawk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_furyhawk_en_5.4.2_3.0_1724078733857.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_furyhawk_en_5.4.2_3.0_1724078733857.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_bbc_furyhawk","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_bbc_furyhawk", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_bbc_furyhawk| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|993.1 MB| + +## References + +https://huggingface.co/furyhawk/t5-base-finetuned-bbc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_pipeline_en.md new file mode 100644 index 00000000000000..b54d0d88b6d5eb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_bbc_furyhawk_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_bbc_furyhawk_pipeline pipeline T5Transformer from furyhawk +author: John Snow Labs +name: t5_base_finetuned_bbc_furyhawk_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_bbc_furyhawk_pipeline` is a English model originally trained by furyhawk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_furyhawk_pipeline_en_5.4.2_3.0_1724078782952.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_bbc_furyhawk_pipeline_en_5.4.2_3.0_1724078782952.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_bbc_furyhawk_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_bbc_furyhawk_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_bbc_furyhawk_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|993.1 MB| + +## References + +https://huggingface.co/furyhawk/t5-base-finetuned-bbc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_en.md new file mode 100644 index 00000000000000..f4448c857b7c28 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_english_tonga_tonga_islands_spanish T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_base_finetuned_english_tonga_tonga_islands_spanish +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_english_tonga_tonga_islands_spanish` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_english_tonga_tonga_islands_spanish_en_5.4.2_3.0_1724053016160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_english_tonga_tonga_islands_spanish_en_5.4.2_3.0_1724053016160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_english_tonga_tonga_islands_spanish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_english_tonga_tonga_islands_spanish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_english_tonga_tonga_islands_spanish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-base-finetuned-en-to-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline_en.md new file mode 100644 index 00000000000000..7051f4c912b7e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline pipeline T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline_en_5.4.2_3.0_1724053062145.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline_en_5.4.2_3.0_1724053062145.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_english_tonga_tonga_islands_spanish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-base-finetuned-en-to-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_en.md new file mode 100644 index 00000000000000..7ed2cb1f798eb1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_ner_2306_1815 T5Transformer from zblaaa +author: John Snow Labs +name: t5_base_finetuned_ner_2306_1815 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_ner_2306_1815` is a English model originally trained by zblaaa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_ner_2306_1815_en_5.4.2_3.0_1724060875535.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_ner_2306_1815_en_5.4.2_3.0_1724060875535.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_ner_2306_1815","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_ner_2306_1815", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_ner_2306_1815| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|957.5 MB| + +## References + +https://huggingface.co/zblaaa/t5-base-finetuned-ner_2306_1815 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_pipeline_en.md new file mode 100644 index 00000000000000..65548991b81853 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_ner_2306_1815_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_ner_2306_1815_pipeline pipeline T5Transformer from zblaaa +author: John Snow Labs +name: t5_base_finetuned_ner_2306_1815_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_ner_2306_1815_pipeline` is a English model originally trained by zblaaa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_ner_2306_1815_pipeline_en_5.4.2_3.0_1724060924169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_ner_2306_1815_pipeline_en_5.4.2_3.0_1724060924169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_ner_2306_1815_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_ner_2306_1815_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_ner_2306_1815_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|957.5 MB| + +## References + +https://huggingface.co/zblaaa/t5-base-finetuned-ner_2306_1815 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_en.md new file mode 100644 index 00000000000000..7cd9430715e469 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_news_summary T5Transformer from radhikabansal +author: John Snow Labs +name: t5_base_finetuned_news_summary +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_news_summary` is a English model originally trained by radhikabansal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_news_summary_en_5.4.2_3.0_1724037704124.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_news_summary_en_5.4.2_3.0_1724037704124.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_news_summary","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_news_summary", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_news_summary| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|974.4 MB| + +## References + +https://huggingface.co/radhikabansal/t5-base-finetuned-news-summary \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_pipeline_en.md new file mode 100644 index 00000000000000..fbffe54434c52a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_news_summary_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_news_summary_pipeline pipeline T5Transformer from radhikabansal +author: John Snow Labs +name: t5_base_finetuned_news_summary_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_news_summary_pipeline` is a English model originally trained by radhikabansal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_news_summary_pipeline_en_5.4.2_3.0_1724037756277.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_news_summary_pipeline_en_5.4.2_3.0_1724037756277.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_news_summary_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_news_summary_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_news_summary_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|974.4 MB| + +## References + +https://huggingface.co/radhikabansal/t5-base-finetuned-news-summary + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_en.md new file mode 100644 index 00000000000000..d0fd59567d2682 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_scitldr_param2 T5Transformer from witchling22 +author: John Snow Labs +name: t5_base_finetuned_scitldr_param2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_scitldr_param2` is a English model originally trained by witchling22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_scitldr_param2_en_5.4.2_3.0_1724097189577.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_scitldr_param2_en_5.4.2_3.0_1724097189577.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_scitldr_param2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_scitldr_param2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_scitldr_param2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|971.1 MB| + +## References + +https://huggingface.co/witchling22/t5-base-finetuned-scitldr_param2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_pipeline_en.md new file mode 100644 index 00000000000000..b3993a10e3d007 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_scitldr_param2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_scitldr_param2_pipeline pipeline T5Transformer from witchling22 +author: John Snow Labs +name: t5_base_finetuned_scitldr_param2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_scitldr_param2_pipeline` is a English model originally trained by witchling22. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_scitldr_param2_pipeline_en_5.4.2_3.0_1724097245979.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_scitldr_param2_pipeline_en_5.4.2_3.0_1724097245979.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_scitldr_param2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_scitldr_param2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_scitldr_param2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|971.1 MB| + +## References + +https://huggingface.co/witchling22/t5-base-finetuned-scitldr_param2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_en.md new file mode 100644 index 00000000000000..85f35e77eaaaec --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_spa_squadv1 T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_spa_squadv1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spa_squadv1` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spa_squadv1_en_5.4.2_3.0_1724026120856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spa_squadv1_en_5.4.2_3.0_1724026120856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_spa_squadv1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_spa_squadv1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spa_squadv1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|954.1 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-spa-squadv1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_pipeline_en.md new file mode 100644 index 00000000000000..8973ce5f3b23b8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_spa_squadv1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_spa_squadv1_pipeline pipeline T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_spa_squadv1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_spa_squadv1_pipeline` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spa_squadv1_pipeline_en_5.4.2_3.0_1724026181266.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_spa_squadv1_pipeline_en_5.4.2_3.0_1724026181266.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_spa_squadv1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_spa_squadv1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_spa_squadv1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|954.1 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-spa-squadv1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_en.md new file mode 100644 index 00000000000000..6060f2bcb4d0c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_stocknews_1900_100_sujayc66 T5Transformer from sujayC66 +author: John Snow Labs +name: t5_base_finetuned_stocknews_1900_100_sujayc66 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_stocknews_1900_100_sujayc66` is a English model originally trained by sujayC66. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_1900_100_sujayc66_en_5.4.2_3.0_1724100057448.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_1900_100_sujayc66_en_5.4.2_3.0_1724100057448.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_stocknews_1900_100_sujayc66","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_stocknews_1900_100_sujayc66", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_stocknews_1900_100_sujayc66| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|989.0 MB| + +## References + +https://huggingface.co/sujayC66/t5-base-finetuned-stocknews_1900_100 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline_en.md new file mode 100644 index 00000000000000..9cfa01d303b902 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline pipeline T5Transformer from sujayC66 +author: John Snow Labs +name: t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline` is a English model originally trained by sujayC66. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline_en_5.4.2_3.0_1724100111796.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline_en_5.4.2_3.0_1724100111796.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_stocknews_1900_100_sujayc66_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|989.0 MB| + +## References + +https://huggingface.co/sujayC66/t5-base-finetuned-stocknews_1900_100 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_en.md new file mode 100644 index 00000000000000..c94506c9f9f3f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_stocknews_2000_150 T5Transformer from sujayC66 +author: John Snow Labs +name: t5_base_finetuned_stocknews_2000_150 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_stocknews_2000_150` is a English model originally trained by sujayC66. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_2000_150_en_5.4.2_3.0_1724067641257.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_2000_150_en_5.4.2_3.0_1724067641257.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_stocknews_2000_150","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_stocknews_2000_150", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_stocknews_2000_150| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.6 MB| + +## References + +https://huggingface.co/sujayC66/t5-base-finetuned-stocknews_2000_150 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_pipeline_en.md new file mode 100644 index 00000000000000..8a65fe7df72863 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_stocknews_2000_150_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_stocknews_2000_150_pipeline pipeline T5Transformer from sujayC66 +author: John Snow Labs +name: t5_base_finetuned_stocknews_2000_150_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_stocknews_2000_150_pipeline` is a English model originally trained by sujayC66. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_2000_150_pipeline_en_5.4.2_3.0_1724067696459.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_stocknews_2000_150_pipeline_en_5.4.2_3.0_1724067696459.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_stocknews_2000_150_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_stocknews_2000_150_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_stocknews_2000_150_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.6 MB| + +## References + +https://huggingface.co/sujayC66/t5-base-finetuned-stocknews_2000_150 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_en.md new file mode 100644 index 00000000000000..5f57969c4bb0f5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_tsetsuuhei T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_base_finetuned_tsetsuuhei +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_tsetsuuhei` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tsetsuuhei_en_5.4.2_3.0_1724043559576.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tsetsuuhei_en_5.4.2_3.0_1724043559576.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_tsetsuuhei","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_tsetsuuhei", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_tsetsuuhei| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-base-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_pipeline_en.md new file mode 100644 index 00000000000000..a64fadbd3a18b1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_tsetsuuhei_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_tsetsuuhei_pipeline pipeline T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_base_finetuned_tsetsuuhei_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_tsetsuuhei_pipeline` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tsetsuuhei_pipeline_en_5.4.2_3.0_1724043608151.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_tsetsuuhei_pipeline_en_5.4.2_3.0_1724043608151.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_tsetsuuhei_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_tsetsuuhei_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_tsetsuuhei_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-base-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_en.md new file mode 100644 index 00000000000000..28b29fa5219ad5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i T5Transformer from anusha +author: John Snow Labs +name: t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i` is a English model originally trained by anusha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_en_5.4.2_3.0_1724058698503.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_en_5.4.2_3.0_1724058698503.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|962.5 MB| + +## References + +https://huggingface.co/anusha/t5-base-finetuned-wikiSQL-sql-to-en_15i \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline_en.md new file mode 100644 index 00000000000000..5476bbe963365e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline pipeline T5Transformer from anusha +author: John Snow Labs +name: t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline` is a English model originally trained by anusha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline_en_5.4.2_3.0_1724058756415.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline_en_5.4.2_3.0_1724058756415.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_wikisql_sql_tonga_tonga_islands_english_15i_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|962.5 MB| + +## References + +https://huggingface.co/anusha/t5-base-finetuned-wikiSQL-sql-to-en_15i + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_en.md new file mode 100644 index 00000000000000..b5608286814f12 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_xsum_brutusxu T5Transformer from brutusxu +author: John Snow Labs +name: t5_base_finetuned_xsum_brutusxu +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_xsum_brutusxu` is a English model originally trained by brutusxu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_xsum_brutusxu_en_5.4.2_3.0_1724076751205.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_xsum_brutusxu_en_5.4.2_3.0_1724076751205.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_xsum_brutusxu","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_xsum_brutusxu", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_xsum_brutusxu| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/brutusxu/t5-base-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_pipeline_en.md new file mode 100644 index 00000000000000..a865035dff4d4a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_finetuned_xsum_brutusxu_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_xsum_brutusxu_pipeline pipeline T5Transformer from brutusxu +author: John Snow Labs +name: t5_base_finetuned_xsum_brutusxu_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_xsum_brutusxu_pipeline` is a English model originally trained by brutusxu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_xsum_brutusxu_pipeline_en_5.4.2_3.0_1724076799586.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_xsum_brutusxu_pipeline_en_5.4.2_3.0_1724076799586.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_xsum_brutusxu_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_xsum_brutusxu_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_xsum_brutusxu_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/brutusxu/t5-base-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_en.md new file mode 100644 index 00000000000000..5c554e864391a5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_generative_qa T5Transformer from ramashisx +author: John Snow Labs +name: t5_base_generative_qa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_generative_qa` is a English model originally trained by ramashisx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_generative_qa_en_5.4.2_3.0_1724071214864.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_generative_qa_en_5.4.2_3.0_1724071214864.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_generative_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_generative_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_generative_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ramashisx/t5_base_generative_qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_pipeline_en.md new file mode 100644 index 00000000000000..c062194662effe --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_generative_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_generative_qa_pipeline pipeline T5Transformer from ramashisx +author: John Snow Labs +name: t5_base_generative_qa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_generative_qa_pipeline` is a English model originally trained by ramashisx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_generative_qa_pipeline_en_5.4.2_3.0_1724071264494.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_generative_qa_pipeline_en_5.4.2_3.0_1724071264494.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_generative_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_generative_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_generative_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ramashisx/t5_base_generative_qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_en.md new file mode 100644 index 00000000000000..278293adcb630c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_moe_ex16 T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_moe_ex16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_moe_ex16` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_moe_ex16_en_5.4.2_3.0_1724105313371.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_moe_ex16_en_5.4.2_3.0_1724105313371.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_moe_ex16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_moe_ex16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_moe_ex16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_moe_ex16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_pipeline_en.md new file mode 100644 index 00000000000000..feaac285511631 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_moe_ex16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_moe_ex16_pipeline pipeline T5Transformer from lukeleeai +author: John Snow Labs +name: t5_base_moe_ex16_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_moe_ex16_pipeline` is a English model originally trained by lukeleeai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_moe_ex16_pipeline_en_5.4.2_3.0_1724105464880.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_moe_ex16_pipeline_en_5.4.2_3.0_1724105464880.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_moe_ex16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_moe_ex16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_moe_ex16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|624.5 MB| + +## References + +https://huggingface.co/lukeleeai/t5-base_moe_ex16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_en.md new file mode 100644 index 00000000000000..ed7fe5447c3ad7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_msmarco_nlgen_ob T5Transformer from din0s +author: John Snow Labs +name: t5_base_msmarco_nlgen_ob +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_msmarco_nlgen_ob` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_msmarco_nlgen_ob_en_5.4.2_3.0_1724061325132.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_msmarco_nlgen_ob_en_5.4.2_3.0_1724061325132.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_msmarco_nlgen_ob","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_msmarco_nlgen_ob", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_msmarco_nlgen_ob| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/din0s/t5-base-msmarco-nlgen-ob \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_pipeline_en.md new file mode 100644 index 00000000000000..5a4de5222f1108 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_msmarco_nlgen_ob_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_msmarco_nlgen_ob_pipeline pipeline T5Transformer from din0s +author: John Snow Labs +name: t5_base_msmarco_nlgen_ob_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_msmarco_nlgen_ob_pipeline` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_msmarco_nlgen_ob_pipeline_en_5.4.2_3.0_1724061372791.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_msmarco_nlgen_ob_pipeline_en_5.4.2_3.0_1724061372791.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_msmarco_nlgen_ob_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_msmarco_nlgen_ob_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_msmarco_nlgen_ob_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/din0s/t5-base-msmarco-nlgen-ob + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_en.md new file mode 100644 index 00000000000000..2f7293193c523e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_multi_english_wiki_news_airklizz T5Transformer from airKlizz +author: John Snow Labs +name: t5_base_multi_english_wiki_news_airklizz +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_multi_english_wiki_news_airklizz` is a English model originally trained by airKlizz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_multi_english_wiki_news_airklizz_en_5.4.2_3.0_1724046899431.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_multi_english_wiki_news_airklizz_en_5.4.2_3.0_1724046899431.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_multi_english_wiki_news_airklizz","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_multi_english_wiki_news_airklizz", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_multi_english_wiki_news_airklizz| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/airKlizz/t5-base-multi-en-wiki-news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_pipeline_en.md new file mode 100644 index 00000000000000..2b331a4e47d346 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_multi_english_wiki_news_airklizz_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_multi_english_wiki_news_airklizz_pipeline pipeline T5Transformer from airKlizz +author: John Snow Labs +name: t5_base_multi_english_wiki_news_airklizz_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_multi_english_wiki_news_airklizz_pipeline` is a English model originally trained by airKlizz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_multi_english_wiki_news_airklizz_pipeline_en_5.4.2_3.0_1724046947186.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_multi_english_wiki_news_airklizz_pipeline_en_5.4.2_3.0_1724046947186.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_multi_english_wiki_news_airklizz_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_multi_english_wiki_news_airklizz_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_multi_english_wiki_news_airklizz_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/airKlizz/t5-base-multi-en-wiki-news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_en.md new file mode 100644 index 00000000000000..ed733cbba15baa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_newsqa_qag_trained T5Transformer from StellarMilk +author: John Snow Labs +name: t5_base_newsqa_qag_trained +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_newsqa_qag_trained` is a English model originally trained by StellarMilk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_newsqa_qag_trained_en_5.4.2_3.0_1724068624196.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_newsqa_qag_trained_en_5.4.2_3.0_1724068624196.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_newsqa_qag_trained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_newsqa_qag_trained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_newsqa_qag_trained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/StellarMilk/t5-base-newsqa-qag-trained \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_pipeline_en.md new file mode 100644 index 00000000000000..35b0abb9ccdb41 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_newsqa_qag_trained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_newsqa_qag_trained_pipeline pipeline T5Transformer from StellarMilk +author: John Snow Labs +name: t5_base_newsqa_qag_trained_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_newsqa_qag_trained_pipeline` is a English model originally trained by StellarMilk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_newsqa_qag_trained_pipeline_en_5.4.2_3.0_1724068671542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_newsqa_qag_trained_pipeline_en_5.4.2_3.0_1724068671542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_newsqa_qag_trained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_newsqa_qag_trained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_newsqa_qag_trained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/StellarMilk/t5-base-newsqa-qag-trained + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_en.md new file mode 100644 index 00000000000000..e152a8425e9e48 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_pointer_cstop_artificial T5Transformer from WillHeld +author: John Snow Labs +name: t5_base_pointer_cstop_artificial +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_pointer_cstop_artificial` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_pointer_cstop_artificial_en_5.4.2_3.0_1724089094555.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_pointer_cstop_artificial_en_5.4.2_3.0_1724089094555.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_pointer_cstop_artificial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_pointer_cstop_artificial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_pointer_cstop_artificial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/WillHeld/t5-base-pointer-cstop_artificial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_pipeline_en.md new file mode 100644 index 00000000000000..18df6bd3dd5792 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_pointer_cstop_artificial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_pointer_cstop_artificial_pipeline pipeline T5Transformer from WillHeld +author: John Snow Labs +name: t5_base_pointer_cstop_artificial_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_pointer_cstop_artificial_pipeline` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_pointer_cstop_artificial_pipeline_en_5.4.2_3.0_1724089501857.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_pointer_cstop_artificial_pipeline_en_5.4.2_3.0_1724089501857.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_pointer_cstop_artificial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_pointer_cstop_artificial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_pointer_cstop_artificial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/WillHeld/t5-base-pointer-cstop_artificial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_en.md new file mode 100644 index 00000000000000..8aca1adbc47dbf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_tfidf_amazon_beauty T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_tfidf_amazon_beauty +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_tfidf_amazon_beauty` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_beauty_en_5.4.2_3.0_1724039816029.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_beauty_en_5.4.2_3.0_1724039816029.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_tfidf_amazon_beauty","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_tfidf_amazon_beauty", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_tfidf_amazon_beauty| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|987.3 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-tfidf-amazon-beauty \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_pipeline_en.md new file mode 100644 index 00000000000000..0e4075acf20d80 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_beauty_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_tfidf_amazon_beauty_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_tfidf_amazon_beauty_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_tfidf_amazon_beauty_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_beauty_pipeline_en_5.4.2_3.0_1724039870827.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_beauty_pipeline_en_5.4.2_3.0_1724039870827.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_tfidf_amazon_beauty_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_tfidf_amazon_beauty_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_tfidf_amazon_beauty_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|987.3 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-tfidf-amazon-beauty + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_en.md new file mode 100644 index 00000000000000..492f7c32798f1c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_rlhf_tfidf_amazon_electronics T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_tfidf_amazon_electronics +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_tfidf_amazon_electronics` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_electronics_en_5.4.2_3.0_1724077089789.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_electronics_en_5.4.2_3.0_1724077089789.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_rlhf_tfidf_amazon_electronics","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_rlhf_tfidf_amazon_electronics", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_tfidf_amazon_electronics| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.6 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-tfidf-amazon-electronics \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_pipeline_en.md new file mode 100644 index 00000000000000..061342a5f40db2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_rlhf_tfidf_amazon_electronics_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_rlhf_tfidf_amazon_electronics_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_rlhf_tfidf_amazon_electronics_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_rlhf_tfidf_amazon_electronics_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_electronics_pipeline_en_5.4.2_3.0_1724077142817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_rlhf_tfidf_amazon_electronics_pipeline_en_5.4.2_3.0_1724077142817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_rlhf_tfidf_amazon_electronics_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_rlhf_tfidf_amazon_electronics_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_rlhf_tfidf_amazon_electronics_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.6 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-rlhf-tfidf-amazon-electronics + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_en.md new file mode 100644 index 00000000000000..a2d101414f3317 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_romanian_finetuned_english_tonga_tonga_islands_italian T5Transformer from din0s +author: John Snow Labs +name: t5_base_romanian_finetuned_english_tonga_tonga_islands_italian +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_romanian_finetuned_english_tonga_tonga_islands_italian` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724086545620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_en_5.4.2_3.0_1724086545620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_romanian_finetuned_english_tonga_tonga_islands_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_romanian_finetuned_english_tonga_tonga_islands_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_romanian_finetuned_english_tonga_tonga_islands_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/din0s/t5-base_ro-finetuned-en-to-it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md new file mode 100644 index 00000000000000..86439e73ae42e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline pipeline T5Transformer from din0s +author: John Snow Labs +name: t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline` is a English model originally trained by din0s. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724086593289.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline_en_5.4.2_3.0_1724086593289.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_romanian_finetuned_english_tonga_tonga_islands_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/din0s/t5-base_ro-finetuned-en-to-it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_en.md new file mode 100644 index 00000000000000..76611b012261c9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_samsum_seed55 T5Transformer from PSW +author: John Snow Labs +name: t5_base_samsum_seed55 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_samsum_seed55` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed55_en_5.4.2_3.0_1724082023095.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed55_en_5.4.2_3.0_1724082023095.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_samsum_seed55","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_samsum_seed55", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_samsum_seed55| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-samsum-seed55 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_pipeline_en.md new file mode 100644 index 00000000000000..193459d22a944c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_samsum_seed55_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_samsum_seed55_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_samsum_seed55_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_samsum_seed55_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed55_pipeline_en_5.4.2_3.0_1724082072987.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed55_pipeline_en_5.4.2_3.0_1724082072987.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_samsum_seed55_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_samsum_seed55_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_samsum_seed55_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-samsum-seed55 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_en.md new file mode 100644 index 00000000000000..5f7a723f7cadc0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sft_amazon_electronics T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_amazon_electronics +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_amazon_electronics` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_amazon_electronics_en_5.4.2_3.0_1724051385223.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_amazon_electronics_en_5.4.2_3.0_1724051385223.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sft_amazon_electronics","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sft_amazon_electronics", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_amazon_electronics| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.3 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-amazon-electronics \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_pipeline_en.md new file mode 100644 index 00000000000000..c4b39039e7bb6d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_amazon_electronics_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sft_amazon_electronics_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_amazon_electronics_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_amazon_electronics_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_amazon_electronics_pipeline_en_5.4.2_3.0_1724051438210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_amazon_electronics_pipeline_en_5.4.2_3.0_1724051438210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sft_amazon_electronics_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sft_amazon_electronics_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_amazon_electronics_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.3 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-amazon-electronics + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_en.md new file mode 100644 index 00000000000000..5b60de3a6b179b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sft_phones T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_phones +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_phones` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_phones_en_5.4.2_3.0_1724060791133.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_phones_en_5.4.2_3.0_1724060791133.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sft_phones","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sft_phones", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_phones| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|987.5 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-phones \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_pipeline_en.md new file mode 100644 index 00000000000000..38a462f61b9ce6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_sft_phones_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sft_phones_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_phones_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_phones_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_phones_pipeline_en_5.4.2_3.0_1724060845446.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_phones_pipeline_en_5.4.2_3.0_1724060845446.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sft_phones_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sft_phones_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_phones_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|987.5 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-phones + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_en.md new file mode 100644 index 00000000000000..d17ffe77f17dfd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_squad_qg_ep10 T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_squad_qg_ep10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squad_qg_ep10` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squad_qg_ep10_en_5.4.2_3.0_1724055375900.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squad_qg_ep10_en_5.4.2_3.0_1724055375900.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_squad_qg_ep10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_squad_qg_ep10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squad_qg_ep10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|993.4 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-SQuAD-qg-ep10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_pipeline_en.md new file mode 100644 index 00000000000000..0dba0f4723c215 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_squad_qg_ep10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_squad_qg_ep10_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: t5_base_squad_qg_ep10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_squad_qg_ep10_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_squad_qg_ep10_pipeline_en_5.4.2_3.0_1724055423637.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_squad_qg_ep10_pipeline_en_5.4.2_3.0_1724055423637.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_squad_qg_ep10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_squad_qg_ep10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_squad_qg_ep10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|993.4 MB| + +## References + +https://huggingface.co/Zekunli/t5-base-SQuAD-qg-ep10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_en.md new file mode 100644 index 00000000000000..4d0144179eb6a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_subjqa_electronics_qg T5Transformer from research-backup +author: John Snow Labs +name: t5_base_subjqa_electronics_qg +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_subjqa_electronics_qg` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_subjqa_electronics_qg_en_5.4.2_3.0_1724063877725.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_subjqa_electronics_qg_en_5.4.2_3.0_1724063877725.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_subjqa_electronics_qg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_subjqa_electronics_qg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_subjqa_electronics_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-subjqa-electronics-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_pipeline_en.md new file mode 100644 index 00000000000000..679c5d56907aed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_subjqa_electronics_qg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_subjqa_electronics_qg_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_base_subjqa_electronics_qg_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_subjqa_electronics_qg_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_subjqa_electronics_qg_pipeline_en_5.4.2_3.0_1724063926284.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_subjqa_electronics_qg_pipeline_en_5.4.2_3.0_1724063926284.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_subjqa_electronics_qg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_subjqa_electronics_qg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_subjqa_electronics_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/t5-base-subjqa-electronics-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_en.md new file mode 100644 index 00000000000000..668403d2bca03a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_summarization_omar47 T5Transformer from omar47 +author: John Snow Labs +name: t5_base_summarization_omar47 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_summarization_omar47` is a English model originally trained by omar47. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_summarization_omar47_en_5.4.2_3.0_1724096364889.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_summarization_omar47_en_5.4.2_3.0_1724096364889.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_summarization_omar47","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_summarization_omar47", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_summarization_omar47| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/omar47/t5-base-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_pipeline_en.md new file mode 100644 index 00000000000000..27602f8c7d69e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_summarization_omar47_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_summarization_omar47_pipeline pipeline T5Transformer from omar47 +author: John Snow Labs +name: t5_base_summarization_omar47_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_summarization_omar47_pipeline` is a English model originally trained by omar47. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_summarization_omar47_pipeline_en_5.4.2_3.0_1724096415743.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_summarization_omar47_pipeline_en_5.4.2_3.0_1724096415743.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_summarization_omar47_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_summarization_omar47_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_summarization_omar47_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/omar47/t5-base-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_en.md new file mode 100644 index 00000000000000..81beed11f30fa7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_0front_1body_7rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_0front_1body_7rear +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_0front_1body_7rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_0front_1body_7rear_en_5.4.2_3.0_1724028059171.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_0front_1body_7rear_en_5.4.2_3.0_1724028059171.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_0front_1body_7rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_0front_1body_7rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_0front_1body_7rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-0front-1body-7rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_pipeline_en.md new file mode 100644 index 00000000000000..f51eae5ea818b9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_0front_1body_7rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_0front_1body_7rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_0front_1body_7rear_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_0front_1body_7rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_0front_1body_7rear_pipeline_en_5.4.2_3.0_1724028106650.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_0front_1body_7rear_pipeline_en_5.4.2_3.0_1724028106650.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_0front_1body_7rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_0front_1body_7rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_0front_1body_7rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-0front-1body-7rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_en.md new file mode 100644 index 00000000000000..371873cd137fce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_1body_2context T5Transformer from tyoyo +author: John Snow Labs +name: t5_base_tedxjp_1body_2context +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_1body_2context` is a English model originally trained by tyoyo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1body_2context_en_5.4.2_3.0_1724046359103.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1body_2context_en_5.4.2_3.0_1724046359103.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_1body_2context","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_1body_2context", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_1body_2context| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tyoyo/t5-base-TEDxJP-1body-2context \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_pipeline_en.md new file mode 100644 index 00000000000000..0d4098c17413fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_1body_2context_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_1body_2context_pipeline pipeline T5Transformer from tyoyo +author: John Snow Labs +name: t5_base_tedxjp_1body_2context_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_1body_2context_pipeline` is a English model originally trained by tyoyo. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1body_2context_pipeline_en_5.4.2_3.0_1724046408497.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1body_2context_pipeline_en_5.4.2_3.0_1724046408497.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_1body_2context_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_1body_2context_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_1body_2context_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tyoyo/t5-base-TEDxJP-1body-2context + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_en.md new file mode 100644 index 00000000000000..9dcbf36a698611 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_9front_1body_9rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_9front_1body_9rear +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_9front_1body_9rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_9rear_en_5.4.2_3.0_1724078141356.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_9rear_en_5.4.2_3.0_1724078141356.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_9front_1body_9rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_9front_1body_9rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_9front_1body_9rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-9front-1body-9rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_pipeline_en.md new file mode 100644 index 00000000000000..d73b7f733f6b49 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tedxjp_9front_1body_9rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_9front_1body_9rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_9front_1body_9rear_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_9front_1body_9rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_9rear_pipeline_en_5.4.2_3.0_1724078193245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_9front_1body_9rear_pipeline_en_5.4.2_3.0_1724078193245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_9front_1body_9rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_9front_1body_9rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_9front_1body_9rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-9front-1body-9rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_en.md new file mode 100644 index 00000000000000..32f416cabd50e6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tweetqa_qa T5Transformer from lmqg +author: John Snow Labs +name: t5_base_tweetqa_qa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tweetqa_qa` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tweetqa_qa_en_5.4.2_3.0_1724063901145.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tweetqa_qa_en_5.4.2_3.0_1724063901145.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tweetqa_qa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tweetqa_qa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tweetqa_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lmqg/t5-base-tweetqa-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_pipeline_en.md new file mode 100644 index 00000000000000..b8792e5166c8ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_tweetqa_qa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tweetqa_qa_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: t5_base_tweetqa_qa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tweetqa_qa_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tweetqa_qa_pipeline_en_5.4.2_3.0_1724063949045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tweetqa_qa_pipeline_en_5.4.2_3.0_1724063949045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tweetqa_qa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tweetqa_qa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tweetqa_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lmqg/t5-base-tweetqa-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_en.md new file mode 100644 index 00000000000000..8f5883aab56154 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_wsc T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_base_wsc +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_wsc` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_wsc_en_5.4.2_3.0_1724042795721.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_wsc_en_5.4.2_3.0_1724042795721.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_wsc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_wsc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_wsc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|976.6 MB| + +## References + +https://huggingface.co/SeongwooKim/T5-base-wsc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_pipeline_en.md new file mode 100644 index 00000000000000..26eac003245b3c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_base_wsc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_wsc_pipeline pipeline T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_base_wsc_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_wsc_pipeline` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_wsc_pipeline_en_5.4.2_3.0_1724042852861.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_wsc_pipeline_en_5.4.2_3.0_1724042852861.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_wsc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_wsc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_wsc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|976.6 MB| + +## References + +https://huggingface.co/SeongwooKim/T5-base-wsc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_en.md new file mode 100644 index 00000000000000..93fa08267c4a59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_boolq_afework T5Transformer from Afework +author: John Snow Labs +name: t5_boolq_afework +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_boolq_afework` is a English model originally trained by Afework. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_boolq_afework_en_5.4.2_3.0_1724101379327.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_boolq_afework_en_5.4.2_3.0_1724101379327.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_boolq_afework","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_boolq_afework", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_boolq_afework| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.1 MB| + +## References + +https://huggingface.co/Afework/t5_boolq \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_pipeline_en.md new file mode 100644 index 00000000000000..a6ae6861824610 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_boolq_afework_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_boolq_afework_pipeline pipeline T5Transformer from Afework +author: John Snow Labs +name: t5_boolq_afework_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_boolq_afework_pipeline` is a English model originally trained by Afework. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_boolq_afework_pipeline_en_5.4.2_3.0_1724101433723.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_boolq_afework_pipeline_en_5.4.2_3.0_1724101433723.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_boolq_afework_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_boolq_afework_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_boolq_afework_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.1 MB| + +## References + +https://huggingface.co/Afework/t5_boolq + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_pipeline_zh.md b/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_pipeline_zh.md new file mode 100644 index 00000000000000..29bc7479031c79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_pipeline_zh.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Chinese t5_chinese_couplet_pipeline pipeline T5Transformer from shibing624 +author: John Snow Labs +name: t5_chinese_couplet_pipeline +date: 2024-08-19 +tags: [zh, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_chinese_couplet_pipeline` is a Chinese model originally trained by shibing624. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_chinese_couplet_pipeline_zh_5.4.2_3.0_1724026313843.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_chinese_couplet_pipeline_zh_5.4.2_3.0_1724026313843.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_chinese_couplet_pipeline", lang = "zh") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_chinese_couplet_pipeline", lang = "zh") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_chinese_couplet_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|zh| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shibing624/t5-chinese-couplet + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_zh.md b/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_zh.md new file mode 100644 index 00000000000000..b0f398d23d9a16 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_chinese_couplet_zh.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Chinese t5_chinese_couplet T5Transformer from shibing624 +author: John Snow Labs +name: t5_chinese_couplet +date: 2024-08-19 +tags: [zh, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_chinese_couplet` is a Chinese model originally trained by shibing624. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_chinese_couplet_zh_5.4.2_3.0_1724026269204.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_chinese_couplet_zh_5.4.2_3.0_1724026269204.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_chinese_couplet","zh") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_chinese_couplet", "zh") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_chinese_couplet| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|zh| +|Size:|1.0 GB| + +## References + +https://huggingface.co/shibing624/t5-chinese-couplet \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_en.md new file mode 100644 index 00000000000000..404cab9e996653 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_conditioned_next_turn T5Transformer from benjaminbeilharz +author: John Snow Labs +name: t5_conditioned_next_turn +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_conditioned_next_turn` is a English model originally trained by benjaminbeilharz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_conditioned_next_turn_en_5.4.2_3.0_1724098493799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_conditioned_next_turn_en_5.4.2_3.0_1724098493799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_conditioned_next_turn","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_conditioned_next_turn", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_conditioned_next_turn| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/benjaminbeilharz/t5-conditioned-next-turn \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_pipeline_en.md new file mode 100644 index 00000000000000..1b21caebf870a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_conditioned_next_turn_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_conditioned_next_turn_pipeline pipeline T5Transformer from benjaminbeilharz +author: John Snow Labs +name: t5_conditioned_next_turn_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_conditioned_next_turn_pipeline` is a English model originally trained by benjaminbeilharz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_conditioned_next_turn_pipeline_en_5.4.2_3.0_1724098547747.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_conditioned_next_turn_pipeline_en_5.4.2_3.0_1724098547747.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_conditioned_next_turn_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_conditioned_next_turn_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_conditioned_next_turn_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/benjaminbeilharz/t5-conditioned-next-turn + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_en.md new file mode 100644 index 00000000000000..d46ac0c326847e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_base_nl32 T5Transformer from google +author: John Snow Labs +name: t5_efficient_base_nl32 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_base_nl32` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_base_nl32_en_5.4.2_3.0_1724043707280.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_base_nl32_en_5.4.2_3.0_1724043707280.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_base_nl32","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_base_nl32", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_base_nl32| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/google/t5-efficient-base-nl32 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_pipeline_en.md new file mode 100644 index 00000000000000..d6a5d1086b76d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_base_nl32_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_base_nl32_pipeline pipeline T5Transformer from google +author: John Snow Labs +name: t5_efficient_base_nl32_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_base_nl32_pipeline` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_base_nl32_pipeline_en_5.4.2_3.0_1724044081278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_base_nl32_pipeline_en_5.4.2_3.0_1724044081278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_base_nl32_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_base_nl32_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_base_nl32_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/google/t5-efficient-base-nl32 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_en.md new file mode 100644 index 00000000000000..a2c1343651583a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_gc4_german_base_nl36_old T5Transformer from GermanT5 +author: John Snow Labs +name: t5_efficient_gc4_german_base_nl36_old +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_gc4_german_base_nl36_old` is a English model originally trained by GermanT5. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_german_base_nl36_old_en_5.4.2_3.0_1724028763101.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_german_base_nl36_old_en_5.4.2_3.0_1724028763101.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_gc4_german_base_nl36_old","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_gc4_german_base_nl36_old", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_gc4_german_base_nl36_old| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/GermanT5/t5-efficient-gc4-german-base-nl36-old \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_pipeline_en.md new file mode 100644 index 00000000000000..bbca6ea4f6d63e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_gc4_german_base_nl36_old_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_gc4_german_base_nl36_old_pipeline pipeline T5Transformer from GermanT5 +author: John Snow Labs +name: t5_efficient_gc4_german_base_nl36_old_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_gc4_german_base_nl36_old_pipeline` is a English model originally trained by GermanT5. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_german_base_nl36_old_pipeline_en_5.4.2_3.0_1724029197163.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_gc4_german_base_nl36_old_pipeline_en_5.4.2_3.0_1724029197163.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_gc4_german_base_nl36_old_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_gc4_german_base_nl36_old_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_gc4_german_base_nl36_old_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/GermanT5/t5-efficient-gc4-german-base-nl36-old + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_dl16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_dl16_en.md new file mode 100644 index 00000000000000..32a2188691be06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_dl16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_large_dl16 T5Transformer from google +author: John Snow Labs +name: t5_efficient_large_dl16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_large_dl16` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_large_dl16_en_5.4.2_3.0_1724084858479.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_large_dl16_en_5.4.2_3.0_1724084858479.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_large_dl16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_large_dl16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_large_dl16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/google/t5-efficient-large-dl16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_en.md new file mode 100644 index 00000000000000..07137da9363213 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_large_el12 T5Transformer from google +author: John Snow Labs +name: t5_efficient_large_el12 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_large_el12` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_large_el12_en_5.4.2_3.0_1724067091866.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_large_el12_en_5.4.2_3.0_1724067091866.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_large_el12","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_large_el12", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_large_el12| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/google/t5-efficient-large-el12 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_pipeline_en.md new file mode 100644 index 00000000000000..d75a3a41519f21 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_el12_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_large_el12_pipeline pipeline T5Transformer from google +author: John Snow Labs +name: t5_efficient_large_el12_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_large_el12_pipeline` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_large_el12_pipeline_en_5.4.2_3.0_1724067556255.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_large_el12_pipeline_en_5.4.2_3.0_1724067556255.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_large_el12_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_large_el12_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_large_el12_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/google/t5-efficient-large-el12 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_en.md new file mode 100644 index 00000000000000..52c7398f22e452 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_large_kv16 T5Transformer from google +author: John Snow Labs +name: t5_efficient_large_kv16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_large_kv16` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_large_kv16_en_5.4.2_3.0_1724058827756.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_large_kv16_en_5.4.2_3.0_1724058827756.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_large_kv16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_large_kv16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_large_kv16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/google/t5-efficient-large-kv16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_pipeline_en.md new file mode 100644 index 00000000000000..c671b172a2f719 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_large_kv16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_large_kv16_pipeline pipeline T5Transformer from google +author: John Snow Labs +name: t5_efficient_large_kv16_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_large_kv16_pipeline` is a English model originally trained by google. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_large_kv16_pipeline_en_5.4.2_3.0_1724059182557.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_large_kv16_pipeline_en_5.4.2_3.0_1724059182557.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_large_kv16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_large_kv16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_large_kv16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/google/t5-efficient-large-kv16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_en.md new file mode 100644 index 00000000000000..5c7e1471a4d9b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_efficient_tiny_nl2_finetuned T5Transformer from abwqr +author: John Snow Labs +name: t5_efficient_tiny_nl2_finetuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_tiny_nl2_finetuned` is a English model originally trained by abwqr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_tiny_nl2_finetuned_en_5.4.2_3.0_1724027319035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_tiny_nl2_finetuned_en_5.4.2_3.0_1724027319035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_efficient_tiny_nl2_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_efficient_tiny_nl2_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_tiny_nl2_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|116.0 MB| + +## References + +https://huggingface.co/abwqr/t5-efficient-tiny-nl2-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..597f70f0b6c649 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_efficient_tiny_nl2_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_efficient_tiny_nl2_finetuned_pipeline pipeline T5Transformer from abwqr +author: John Snow Labs +name: t5_efficient_tiny_nl2_finetuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_efficient_tiny_nl2_finetuned_pipeline` is a English model originally trained by abwqr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_efficient_tiny_nl2_finetuned_pipeline_en_5.4.2_3.0_1724027325155.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_efficient_tiny_nl2_finetuned_pipeline_en_5.4.2_3.0_1724027325155.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_efficient_tiny_nl2_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_efficient_tiny_nl2_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_efficient_tiny_nl2_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|116.0 MB| + +## References + +https://huggingface.co/abwqr/t5-efficient-tiny-nl2-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_en.md new file mode 100644 index 00000000000000..0a63c69886651e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_english_korean_model T5Transformer from halee9 +author: John Snow Labs +name: t5_english_korean_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_english_korean_model` is a English model originally trained by halee9. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_english_korean_model_en_5.4.2_3.0_1724093504869.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_english_korean_model_en_5.4.2_3.0_1724093504869.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_english_korean_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_english_korean_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_english_korean_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|663.5 MB| + +## References + +https://huggingface.co/halee9/t5_en_ko_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_pipeline_en.md new file mode 100644 index 00000000000000..c498dedf3eda94 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_english_korean_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_english_korean_model_pipeline pipeline T5Transformer from halee9 +author: John Snow Labs +name: t5_english_korean_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_english_korean_model_pipeline` is a English model originally trained by halee9. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_english_korean_model_pipeline_en_5.4.2_3.0_1724093733828.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_english_korean_model_pipeline_en_5.4.2_3.0_1724093733828.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_english_korean_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_english_korean_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_english_korean_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|663.5 MB| + +## References + +https://huggingface.co/halee9/t5_en_ko_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_en.md new file mode 100644 index 00000000000000..6c0edc4735b5df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_english_tonga_tonga_islands_japanese_eval1 T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_finetuned_english_tonga_tonga_islands_japanese_eval1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_english_tonga_tonga_islands_japanese_eval1` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_japanese_eval1_en_5.4.2_3.0_1724107443645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_japanese_eval1_en_5.4.2_3.0_1724107443645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_english_tonga_tonga_islands_japanese_eval1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_english_tonga_tonga_islands_japanese_eval1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_english_tonga_tonga_islands_japanese_eval1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-finetuned-en-to-ja-eval1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline_en.md new file mode 100644 index 00000000000000..7f91d1a90699ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline pipeline T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline_en_5.4.2_3.0_1724107516278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline_en_5.4.2_3.0_1724107516278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_english_tonga_tonga_islands_japanese_eval1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-finetuned-en-to-ja-eval1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_en.md new file mode 100644 index 00000000000000..a09d5e03349e9b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_english_tonga_tonga_islands_russian_eval1 T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_finetuned_english_tonga_tonga_islands_russian_eval1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_english_tonga_tonga_islands_russian_eval1` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_russian_eval1_en_5.4.2_3.0_1724111592314.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_russian_eval1_en_5.4.2_3.0_1724111592314.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_english_tonga_tonga_islands_russian_eval1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_english_tonga_tonga_islands_russian_eval1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_english_tonga_tonga_islands_russian_eval1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-finetuned-en-to-ru-eval1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline_en.md new file mode 100644 index 00000000000000..4fd33cda464dae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline pipeline T5Transformer from tsetsuuhei +author: John Snow Labs +name: t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline` is a English model originally trained by tsetsuuhei. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline_en_5.4.2_3.0_1724111643097.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline_en_5.4.2_3.0_1724111643097.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_english_tonga_tonga_islands_russian_eval1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/tsetsuuhei/t5-finetuned-en-to-ru-eval1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_en.md new file mode 100644 index 00000000000000..1a34c7df876d25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_finetuned_paraphrase T5Transformer from Ujjawal +author: John Snow Labs +name: t5_finetuned_paraphrase +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_paraphrase` is a English model originally trained by Ujjawal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_paraphrase_en_5.4.2_3.0_1724084246747.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_paraphrase_en_5.4.2_3.0_1724084246747.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_finetuned_paraphrase","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_finetuned_paraphrase", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_paraphrase| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.5 MB| + +## References + +https://huggingface.co/Ujjawal/t5_finetuned_paraphrase \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_pipeline_en.md new file mode 100644 index 00000000000000..e8ac653749c1d1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_finetuned_paraphrase_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_finetuned_paraphrase_pipeline pipeline T5Transformer from Ujjawal +author: John Snow Labs +name: t5_finetuned_paraphrase_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_finetuned_paraphrase_pipeline` is a English model originally trained by Ujjawal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_finetuned_paraphrase_pipeline_en_5.4.2_3.0_1724084270630.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_finetuned_paraphrase_pipeline_en_5.4.2_3.0_1724084270630.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_finetuned_paraphrase_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_finetuned_paraphrase_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_finetuned_paraphrase_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.5 MB| + +## References + +https://huggingface.co/Ujjawal/t5_finetuned_paraphrase + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_en.md new file mode 100644 index 00000000000000..142e04c0e2c69c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_flan_large_translation T5Transformer from guyhadad01 +author: John Snow Labs +name: t5_flan_large_translation +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_large_translation` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_large_translation_en_5.4.2_3.0_1724041237263.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_large_translation_en_5.4.2_3.0_1724041237263.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_flan_large_translation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_flan_large_translation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_large_translation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/guyhadad01/t5-flan-large-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_pipeline_en.md new file mode 100644 index 00000000000000..70c865579d5bab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_large_translation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_flan_large_translation_pipeline pipeline T5Transformer from guyhadad01 +author: John Snow Labs +name: t5_flan_large_translation_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_large_translation_pipeline` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_large_translation_pipeline_en_5.4.2_3.0_1724041385999.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_large_translation_pipeline_en_5.4.2_3.0_1724041385999.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_flan_large_translation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_flan_large_translation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_large_translation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/guyhadad01/t5-flan-large-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_en.md new file mode 100644 index 00000000000000..a9a897c8ee8b79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_flan_small_dslack T5Transformer from dslack +author: John Snow Labs +name: t5_flan_small_dslack +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_small_dslack` is a English model originally trained by dslack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_small_dslack_en_5.4.2_3.0_1724029128431.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_small_dslack_en_5.4.2_3.0_1724029128431.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_flan_small_dslack","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_flan_small_dslack", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_small_dslack| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/dslack/t5-flan-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_pipeline_en.md new file mode 100644 index 00000000000000..2af1917b969daf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_flan_small_dslack_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_flan_small_dslack_pipeline pipeline T5Transformer from dslack +author: John Snow Labs +name: t5_flan_small_dslack_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_flan_small_dslack_pipeline` is a English model originally trained by dslack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_flan_small_dslack_pipeline_en_5.4.2_3.0_1724029144991.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_flan_small_dslack_pipeline_en_5.4.2_3.0_1724029144991.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_flan_small_dslack_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_flan_small_dslack_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_flan_small_dslack_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/dslack/t5-flan-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_en.md new file mode 100644 index 00000000000000..3ba96a3a350989 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_gec_128len_9e T5Transformer from anujraymajhi +author: John Snow Labs +name: t5_gec_128len_9e +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec_128len_9e` is a English model originally trained by anujraymajhi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_128len_9e_en_5.4.2_3.0_1724054653742.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_128len_9e_en_5.4.2_3.0_1724054653742.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_gec_128len_9e","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_gec_128len_9e", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec_128len_9e| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|997.5 MB| + +## References + +https://huggingface.co/anujraymajhi/t5-GEC-128len-9e \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_pipeline_en.md new file mode 100644 index 00000000000000..dcff91ee3d35e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_128len_9e_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_gec_128len_9e_pipeline pipeline T5Transformer from anujraymajhi +author: John Snow Labs +name: t5_gec_128len_9e_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec_128len_9e_pipeline` is a English model originally trained by anujraymajhi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_128len_9e_pipeline_en_5.4.2_3.0_1724054701007.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_128len_9e_pipeline_en_5.4.2_3.0_1724054701007.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_gec_128len_9e_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_gec_128len_9e_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec_128len_9e_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|997.5 MB| + +## References + +https://huggingface.co/anujraymajhi/t5-GEC-128len-9e + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_gec_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_en.md new file mode 100644 index 00000000000000..ee381e234d5b18 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_gec T5Transformer from BossLee +author: John Snow Labs +name: t5_gec +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec` is a English model originally trained by BossLee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_en_5.4.2_3.0_1724026008313.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_en_5.4.2_3.0_1724026008313.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_gec","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_gec", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/BossLee/t5-gec \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_gec_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_pipeline_en.md new file mode 100644 index 00000000000000..9ce0fa647795d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_gec_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_gec_pipeline pipeline T5Transformer from BossLee +author: John Snow Labs +name: t5_gec_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_gec_pipeline` is a English model originally trained by BossLee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_gec_pipeline_en_5.4.2_3.0_1724026054987.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_gec_pipeline_en_5.4.2_3.0_1724026054987.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_gec_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_gec_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_gec_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/BossLee/t5-gec + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_en.md new file mode 100644 index 00000000000000..7d42ca9408c4a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_drop T5Transformer from StonyBrookNLP +author: John Snow Labs +name: t5_large_drop +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_drop` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_drop_en_5.4.2_3.0_1724091732528.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_drop_en_5.4.2_3.0_1724091732528.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_drop","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_drop", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_drop| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/StonyBrookNLP/t5-large-drop \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_pipeline_en.md new file mode 100644 index 00000000000000..384b2949172637 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_drop_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_drop_pipeline pipeline T5Transformer from StonyBrookNLP +author: John Snow Labs +name: t5_large_drop_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_drop_pipeline` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_drop_pipeline_en_5.4.2_3.0_1724091877361.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_drop_pipeline_en_5.4.2_3.0_1724091877361.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_drop_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_drop_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_drop_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/StonyBrookNLP/t5-large-drop + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_en.md new file mode 100644 index 00000000000000..01459a81dc6355 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_finetuned_bias_2e10ce74 T5Transformer from ajsmith201 +author: John Snow Labs +name: t5_large_finetuned_bias_2e10ce74 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_finetuned_bias_2e10ce74` is a English model originally trained by ajsmith201. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_bias_2e10ce74_en_5.4.2_3.0_1724039355290.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_bias_2e10ce74_en_5.4.2_3.0_1724039355290.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_finetuned_bias_2e10ce74","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_finetuned_bias_2e10ce74", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_finetuned_bias_2e10ce74| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ajsmith201/t5-large-finetuned-bias-2e10ce74 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_pipeline_en.md new file mode 100644 index 00000000000000..5e912a21cb633b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_finetuned_bias_2e10ce74_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_finetuned_bias_2e10ce74_pipeline pipeline T5Transformer from ajsmith201 +author: John Snow Labs +name: t5_large_finetuned_bias_2e10ce74_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_finetuned_bias_2e10ce74_pipeline` is a English model originally trained by ajsmith201. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_bias_2e10ce74_pipeline_en_5.4.2_3.0_1724039495392.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_finetuned_bias_2e10ce74_pipeline_en_5.4.2_3.0_1724039495392.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_finetuned_bias_2e10ce74_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_finetuned_bias_2e10ce74_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_finetuned_bias_2e10ce74_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ajsmith201/t5-large-finetuned-bias-2e10ce74 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_en.md new file mode 100644 index 00000000000000..dd9ab609a341c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_hoax_fulltext_classifier_1h2r T5Transformer from research-dump +author: John Snow Labs +name: t5_large_hoax_fulltext_classifier_1h2r +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_hoax_fulltext_classifier_1h2r` is a English model originally trained by research-dump. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_hoax_fulltext_classifier_1h2r_en_5.4.2_3.0_1724032582339.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_hoax_fulltext_classifier_1h2r_en_5.4.2_3.0_1724032582339.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_hoax_fulltext_classifier_1h2r","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_hoax_fulltext_classifier_1h2r", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_hoax_fulltext_classifier_1h2r| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/research-dump/t5-large_hoax_fulltext_classifier_1h2r \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_pipeline_en.md new file mode 100644 index 00000000000000..c82d4d8bf14f03 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_hoax_fulltext_classifier_1h2r_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_hoax_fulltext_classifier_1h2r_pipeline pipeline T5Transformer from research-dump +author: John Snow Labs +name: t5_large_hoax_fulltext_classifier_1h2r_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_hoax_fulltext_classifier_1h2r_pipeline` is a English model originally trained by research-dump. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_hoax_fulltext_classifier_1h2r_pipeline_en_5.4.2_3.0_1724032747289.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_hoax_fulltext_classifier_1h2r_pipeline_en_5.4.2_3.0_1724032747289.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_hoax_fulltext_classifier_1h2r_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_hoax_fulltext_classifier_1h2r_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_hoax_fulltext_classifier_1h2r_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/research-dump/t5-large_hoax_fulltext_classifier_1h2r + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_en.md new file mode 100644 index 00000000000000..4ce795350984c5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_qasc T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_large_qasc +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_qasc` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_qasc_en_5.4.2_3.0_1724099101440.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_qasc_en_5.4.2_3.0_1724099101440.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_qasc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_qasc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_qasc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/SeongwooKim/T5-large-qasc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_pipeline_en.md new file mode 100644 index 00000000000000..bc40793aed6637 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_qasc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_qasc_pipeline pipeline T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_large_qasc_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_qasc_pipeline` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_qasc_pipeline_en_5.4.2_3.0_1724099258507.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_qasc_pipeline_en_5.4.2_3.0_1724099258507.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_qasc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_qasc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_qasc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/SeongwooKim/T5-large-qasc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_en.md new file mode 100644 index 00000000000000..98c989e3ec2923 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_radiology_ai_imagingcancer_0_8 T5Transformer from hisaoka +author: John Snow Labs +name: t5_large_radiology_ai_imagingcancer_0_8 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_radiology_ai_imagingcancer_0_8` is a English model originally trained by hisaoka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_8_en_5.4.2_3.0_1724031285820.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_8_en_5.4.2_3.0_1724031285820.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_radiology_ai_imagingcancer_0_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_radiology_ai_imagingcancer_0_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_radiology_ai_imagingcancer_0_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/hisaoka/t5-large_radiology-ai-imagingcancer-0.8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_pipeline_en.md new file mode 100644 index 00000000000000..144f48c5b298ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_radiology_ai_imagingcancer_0_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_radiology_ai_imagingcancer_0_8_pipeline pipeline T5Transformer from hisaoka +author: John Snow Labs +name: t5_large_radiology_ai_imagingcancer_0_8_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_radiology_ai_imagingcancer_0_8_pipeline` is a English model originally trained by hisaoka. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_8_pipeline_en_5.4.2_3.0_1724031436684.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_radiology_ai_imagingcancer_0_8_pipeline_en_5.4.2_3.0_1724031436684.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_radiology_ai_imagingcancer_0_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_radiology_ai_imagingcancer_0_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_radiology_ai_imagingcancer_0_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/hisaoka/t5-large_radiology-ai-imagingcancer-0.8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_en.md new file mode 100644 index 00000000000000..afc99895afd67b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_squad_qg_norwegian_answer T5Transformer from research-backup +author: John Snow Labs +name: t5_large_squad_qg_norwegian_answer +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_squad_qg_norwegian_answer` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_squad_qg_norwegian_answer_en_5.4.2_3.0_1724070558698.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_squad_qg_norwegian_answer_en_5.4.2_3.0_1724070558698.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_squad_qg_norwegian_answer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_squad_qg_norwegian_answer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_squad_qg_norwegian_answer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/research-backup/t5-large-squad-qg-no-answer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_pipeline_en.md new file mode 100644 index 00000000000000..4f41b472c5856d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_large_squad_qg_norwegian_answer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_squad_qg_norwegian_answer_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_large_squad_qg_norwegian_answer_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_squad_qg_norwegian_answer_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_squad_qg_norwegian_answer_pipeline_en_5.4.2_3.0_1724070697619.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_squad_qg_norwegian_answer_pipeline_en_5.4.2_3.0_1724070697619.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_squad_qg_norwegian_answer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_squad_qg_norwegian_answer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_squad_qg_norwegian_answer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/research-backup/t5-large-squad-qg-no-answer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_en.md new file mode 100644 index 00000000000000..80ee0d8ae18561 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_liputan6_finetuned_indonesia_summarization T5Transformer from Radvian +author: John Snow Labs +name: t5_liputan6_finetuned_indonesia_summarization +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_liputan6_finetuned_indonesia_summarization` is a English model originally trained by Radvian. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_liputan6_finetuned_indonesia_summarization_en_5.4.2_3.0_1724086470368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_liputan6_finetuned_indonesia_summarization_en_5.4.2_3.0_1724086470368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_liputan6_finetuned_indonesia_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_liputan6_finetuned_indonesia_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_liputan6_finetuned_indonesia_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/Radvian/t5_liputan6_finetuned_indonesia_summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_pipeline_en.md new file mode 100644 index 00000000000000..656293c3f15c7b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_liputan6_finetuned_indonesia_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_liputan6_finetuned_indonesia_summarization_pipeline pipeline T5Transformer from Radvian +author: John Snow Labs +name: t5_liputan6_finetuned_indonesia_summarization_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_liputan6_finetuned_indonesia_summarization_pipeline` is a English model originally trained by Radvian. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_liputan6_finetuned_indonesia_summarization_pipeline_en_5.4.2_3.0_1724086486775.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_liputan6_finetuned_indonesia_summarization_pipeline_en_5.4.2_3.0_1724086486775.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_liputan6_finetuned_indonesia_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_liputan6_finetuned_indonesia_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_liputan6_finetuned_indonesia_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/Radvian/t5_liputan6_finetuned_indonesia_summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_en.md new file mode 100644 index 00000000000000..2e9ce44e7ddeba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_mcq_question_generator_val T5Transformer from Bilkies +author: John Snow Labs +name: t5_mcq_question_generator_val +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_mcq_question_generator_val` is a English model originally trained by Bilkies. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_mcq_question_generator_val_en_5.4.2_3.0_1724068005958.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_mcq_question_generator_val_en_5.4.2_3.0_1724068005958.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_mcq_question_generator_val","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_mcq_question_generator_val", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_mcq_question_generator_val| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Bilkies/t5-MCQ-question-generator_val \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_pipeline_en.md new file mode 100644 index 00000000000000..0f5de31a14d275 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_mcq_question_generator_val_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_mcq_question_generator_val_pipeline pipeline T5Transformer from Bilkies +author: John Snow Labs +name: t5_mcq_question_generator_val_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_mcq_question_generator_val_pipeline` is a English model originally trained by Bilkies. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_mcq_question_generator_val_pipeline_en_5.4.2_3.0_1724068054938.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_mcq_question_generator_val_pipeline_en_5.4.2_3.0_1724068054938.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_mcq_question_generator_val_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_mcq_question_generator_val_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_mcq_question_generator_val_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Bilkies/t5-MCQ-question-generator_val + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_en.md new file mode 100644 index 00000000000000..6418932c3a824c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_model_1_feedback_2010_e4 T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_feedback_2010_e4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_feedback_2010_e4` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_2010_e4_en_5.4.2_3.0_1724075215104.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_2010_e4_en_5.4.2_3.0_1724075215104.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_model_1_feedback_2010_e4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_model_1_feedback_2010_e4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_feedback_2010_e4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|980.9 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-feedback-2010-e4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_pipeline_en.md new file mode 100644 index 00000000000000..4622581f0cb891 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_model_1_feedback_2010_e4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_model_1_feedback_2010_e4_pipeline pipeline T5Transformer from theojolliffe +author: John Snow Labs +name: t5_model_1_feedback_2010_e4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_model_1_feedback_2010_e4_pipeline` is a English model originally trained by theojolliffe. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_2010_e4_pipeline_en_5.4.2_3.0_1724075265683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_model_1_feedback_2010_e4_pipeline_en_5.4.2_3.0_1724075265683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_model_1_feedback_2010_e4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_model_1_feedback_2010_e4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_model_1_feedback_2010_e4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|980.9 MB| + +## References + +https://huggingface.co/theojolliffe/T5-model-1-feedback-2010-e4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_en.md new file mode 100644 index 00000000000000..5cbee79fd9dbae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_news_summarization T5Transformer from Vasanth +author: John Snow Labs +name: t5_news_summarization +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_news_summarization` is a English model originally trained by Vasanth. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_news_summarization_en_5.4.2_3.0_1724074357636.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_news_summarization_en_5.4.2_3.0_1724074357636.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_news_summarization","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_news_summarization", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_news_summarization| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Vasanth/t5-news-summarization \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_pipeline_en.md new file mode 100644 index 00000000000000..cb5d4e04faa9f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_news_summarization_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_news_summarization_pipeline pipeline T5Transformer from Vasanth +author: John Snow Labs +name: t5_news_summarization_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_news_summarization_pipeline` is a English model originally trained by Vasanth. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_news_summarization_pipeline_en_5.4.2_3.0_1724074411643.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_news_summarization_pipeline_en_5.4.2_3.0_1724074411643.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_news_summarization_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_news_summarization_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_news_summarization_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Vasanth/t5-news-summarization + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_ost_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_ost_en.md new file mode 100644 index 00000000000000..d942cad923a7f6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_ost_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_ost T5Transformer from tkuye +author: John Snow Labs +name: t5_ost +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ost` is a English model originally trained by tkuye. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ost_en_5.4.2_3.0_1724042020992.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ost_en_5.4.2_3.0_1724042020992.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_ost","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_ost", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ost| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|996.1 MB| + +## References + +https://huggingface.co/tkuye/t5-ost \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_ost_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_ost_pipeline_en.md new file mode 100644 index 00000000000000..9b60a2a84e42ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_ost_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_ost_pipeline pipeline T5Transformer from tkuye +author: John Snow Labs +name: t5_ost_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ost_pipeline` is a English model originally trained by tkuye. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ost_pipeline_en_5.4.2_3.0_1724042067559.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ost_pipeline_en_5.4.2_3.0_1724042067559.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_ost_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_ost_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ost_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|996.1 MB| + +## References + +https://huggingface.co/tkuye/t5-ost + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_en.md new file mode 100644 index 00000000000000..2d934ad1687b36 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_patent_translation_korean_english T5Transformer from Chaeyeon +author: John Snow Labs +name: t5_patent_translation_korean_english +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_patent_translation_korean_english` is a English model originally trained by Chaeyeon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_patent_translation_korean_english_en_5.4.2_3.0_1724098449857.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_patent_translation_korean_english_en_5.4.2_3.0_1724098449857.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_patent_translation_korean_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_patent_translation_korean_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_patent_translation_korean_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Chaeyeon/T5-Patent-Translation-ko-en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_pipeline_en.md new file mode 100644 index 00000000000000..fbf765506244fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_patent_translation_korean_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_patent_translation_korean_english_pipeline pipeline T5Transformer from Chaeyeon +author: John Snow Labs +name: t5_patent_translation_korean_english_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_patent_translation_korean_english_pipeline` is a English model originally trained by Chaeyeon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_patent_translation_korean_english_pipeline_en_5.4.2_3.0_1724098510750.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_patent_translation_korean_english_pipeline_en_5.4.2_3.0_1724098510750.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_patent_translation_korean_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_patent_translation_korean_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_patent_translation_korean_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Chaeyeon/T5-Patent-Translation-ko-en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_en.md new file mode 100644 index 00000000000000..d90b8d6ee15b38 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_pretrain T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_en_5.4.2_3.0_1724062179952.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_en_5.4.2_3.0_1724062179952.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_pretrain","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_pretrain", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_pretrain \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_en.md new file mode 100644 index 00000000000000..eed46974ee829f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_pretrain_final_final_final_kaggle_final T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain_final_final_final_kaggle_final +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain_final_final_final_kaggle_final` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_final_final_final_kaggle_final_en_5.4.2_3.0_1724105308512.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_final_final_final_kaggle_final_en_5.4.2_3.0_1724105308512.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_pretrain_final_final_final_kaggle_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_pretrain_final_final_final_kaggle_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain_final_final_final_kaggle_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_pretrain_final_final_final_kaggle_final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_pipeline_en.md new file mode 100644 index 00000000000000..8b4e4cc6ba6cd5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_final_final_final_kaggle_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_pretrain_final_final_final_kaggle_final_pipeline pipeline T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain_final_final_final_kaggle_final_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain_final_final_final_kaggle_final_pipeline` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_final_final_final_kaggle_final_pipeline_en_5.4.2_3.0_1724105325679.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_final_final_final_kaggle_final_pipeline_en_5.4.2_3.0_1724105325679.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_pretrain_final_final_final_kaggle_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_pretrain_final_final_final_kaggle_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain_final_final_final_kaggle_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.6 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_pretrain_final_final_final_kaggle_final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_pipeline_en.md new file mode 100644 index 00000000000000..ef04ac21dc9c9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_pretrain_pipeline pipeline T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain_pipeline` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_pipeline_en_5.4.2_3.0_1724062196855.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_pipeline_en_5.4.2_3.0_1724062196855.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_pretrain_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_pretrain_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5_pretrain + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_en.md new file mode 100644 index 00000000000000..9fd73cab87fe70 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_pretrain_wikipedia_sample_final_final T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain_wikipedia_sample_final_final +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain_wikipedia_sample_final_final` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_wikipedia_sample_final_final_en_5.4.2_3.0_1724075379439.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_wikipedia_sample_final_final_en_5.4.2_3.0_1724075379439.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_pretrain_wikipedia_sample_final_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_pretrain_wikipedia_sample_final_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain_wikipedia_sample_final_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5-pretrain-wikipedia-sample-final-final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_pipeline_en.md new file mode 100644 index 00000000000000..fd4e30f2d66a52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_pretrain_wikipedia_sample_final_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_pretrain_wikipedia_sample_final_final_pipeline pipeline T5Transformer from OmarHaroon01 +author: John Snow Labs +name: t5_pretrain_wikipedia_sample_final_final_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_pretrain_wikipedia_sample_final_final_pipeline` is a English model originally trained by OmarHaroon01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_pretrain_wikipedia_sample_final_final_pipeline_en_5.4.2_3.0_1724075396806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_pretrain_wikipedia_sample_final_final_pipeline_en_5.4.2_3.0_1724075396806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_pretrain_wikipedia_sample_final_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_pretrain_wikipedia_sample_final_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_pretrain_wikipedia_sample_final_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/OmarHaroon01/t5-pretrain-wikipedia-sample-final-final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_en.md new file mode 100644 index 00000000000000..637a380b9ea32a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_recommendation_sports_equipment_english_subhamiiita T5Transformer from subhamiiita +author: John Snow Labs +name: t5_recommendation_sports_equipment_english_subhamiiita +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_recommendation_sports_equipment_english_subhamiiita` is a English model originally trained by subhamiiita. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_subhamiiita_en_5.4.2_3.0_1724105745012.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_subhamiiita_en_5.4.2_3.0_1724105745012.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_recommendation_sports_equipment_english_subhamiiita","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_recommendation_sports_equipment_english_subhamiiita", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_recommendation_sports_equipment_english_subhamiiita| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.7 GB| + +## References + +https://huggingface.co/subhamiiita/t5_recommendation_sports_equipment_english \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_pipeline_en.md new file mode 100644 index 00000000000000..41457e797da64e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_recommendation_sports_equipment_english_subhamiiita_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_recommendation_sports_equipment_english_subhamiiita_pipeline pipeline T5Transformer from subhamiiita +author: John Snow Labs +name: t5_recommendation_sports_equipment_english_subhamiiita_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_recommendation_sports_equipment_english_subhamiiita_pipeline` is a English model originally trained by subhamiiita. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_subhamiiita_pipeline_en_5.4.2_3.0_1724105945476.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_recommendation_sports_equipment_english_subhamiiita_pipeline_en_5.4.2_3.0_1724105945476.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_recommendation_sports_equipment_english_subhamiiita_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_recommendation_sports_equipment_english_subhamiiita_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_recommendation_sports_equipment_english_subhamiiita_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.7 GB| + +## References + +https://huggingface.co/subhamiiita/t5_recommendation_sports_equipment_english + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_en.md new file mode 100644 index 00000000000000..644708fca608d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_simple_adafactor T5Transformer from LadyShizu +author: John Snow Labs +name: t5_simple_adafactor +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_simple_adafactor` is a English model originally trained by LadyShizu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_simple_adafactor_en_5.4.2_3.0_1724103183161.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_simple_adafactor_en_5.4.2_3.0_1724103183161.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_simple_adafactor","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_simple_adafactor", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_simple_adafactor| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LadyShizu/T5_simple_adafactor \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_pipeline_en.md new file mode 100644 index 00000000000000..bf21f351c7b8ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_simple_adafactor_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_simple_adafactor_pipeline pipeline T5Transformer from LadyShizu +author: John Snow Labs +name: t5_simple_adafactor_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_simple_adafactor_pipeline` is a English model originally trained by LadyShizu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_simple_adafactor_pipeline_en_5.4.2_3.0_1724103230195.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_simple_adafactor_pipeline_en_5.4.2_3.0_1724103230195.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_simple_adafactor_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_simple_adafactor_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_simple_adafactor_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/LadyShizu/T5_simple_adafactor + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_en.md new file mode 100644 index 00000000000000..d6e91774c93939 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_slovene_large_oasst_context T5Transformer from vh-student +author: John Snow Labs +name: t5_slovene_large_oasst_context +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_slovene_large_oasst_context` is a English model originally trained by vh-student. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_context_en_5.4.2_3.0_1724084157618.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_context_en_5.4.2_3.0_1724084157618.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_slovene_large_oasst_context","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_slovene_large_oasst_context", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_slovene_large_oasst_context| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/vh-student/t5-sl-large-oasst-context \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_pipeline_en.md new file mode 100644 index 00000000000000..43c17395b0aef1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_slovene_large_oasst_context_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_slovene_large_oasst_context_pipeline pipeline T5Transformer from vh-student +author: John Snow Labs +name: t5_slovene_large_oasst_context_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_slovene_large_oasst_context_pipeline` is a English model originally trained by vh-student. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_context_pipeline_en_5.4.2_3.0_1724084327326.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_context_pipeline_en_5.4.2_3.0_1724084327326.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_slovene_large_oasst_context_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_slovene_large_oasst_context_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_slovene_large_oasst_context_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/vh-student/t5-sl-large-oasst-context + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_en.md new file mode 100644 index 00000000000000..49a9f53cd18f87 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_eurlexsum T5Transformer from Pierre-Arthur +author: John Snow Labs +name: t5_small_eurlexsum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_eurlexsum` is a English model originally trained by Pierre-Arthur. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_eurlexsum_en_5.4.2_3.0_1724059399722.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_eurlexsum_en_5.4.2_3.0_1724059399722.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_eurlexsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_eurlexsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_eurlexsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|296.1 MB| + +## References + +https://huggingface.co/Pierre-Arthur/T5_small_eurlexsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_pipeline_en.md new file mode 100644 index 00000000000000..1ac4409e88be06 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_eurlexsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_eurlexsum_pipeline pipeline T5Transformer from Pierre-Arthur +author: John Snow Labs +name: t5_small_eurlexsum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_eurlexsum_pipeline` is a English model originally trained by Pierre-Arthur. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_eurlexsum_pipeline_en_5.4.2_3.0_1724059420860.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_eurlexsum_pipeline_en_5.4.2_3.0_1724059420860.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_eurlexsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_eurlexsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_eurlexsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|296.1 MB| + +## References + +https://huggingface.co/Pierre-Arthur/T5_small_eurlexsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_en.md new file mode 100644 index 00000000000000..bf7f40dfb52df5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_few_shot_k_16_finetuned_squad_seed_4 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_16_finetuned_squad_seed_4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_16_finetuned_squad_seed_4` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_4_en_5.4.2_3.0_1724092630559.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_4_en_5.4.2_3.0_1724092630559.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_few_shot_k_16_finetuned_squad_seed_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_few_shot_k_16_finetuned_squad_seed_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_16_finetuned_squad_seed_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|292.7 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-16-finetuned-squad-seed-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline_en.md new file mode 100644 index 00000000000000..3b9b3cc7c5e202 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline_en_5.4.2_3.0_1724092663458.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline_en_5.4.2_3.0_1724092663458.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_16_finetuned_squad_seed_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|292.7 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-16-finetuned-squad-seed-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_en.md new file mode 100644 index 00000000000000..f692c5405e460e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned8_english_tonga_tonga_islands_german T5Transformer from alexrfelicio +author: John Snow Labs +name: t5_small_finetuned8_english_tonga_tonga_islands_german +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned8_english_tonga_tonga_islands_german` is a English model originally trained by alexrfelicio. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned8_english_tonga_tonga_islands_german_en_5.4.2_3.0_1724072183799.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned8_english_tonga_tonga_islands_german_en_5.4.2_3.0_1724072183799.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned8_english_tonga_tonga_islands_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned8_english_tonga_tonga_islands_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned8_english_tonga_tonga_islands_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|310.1 MB| + +## References + +https://huggingface.co/alexrfelicio/t5-small-finetuned8-en-to-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline_en.md new file mode 100644 index 00000000000000..4905c2267fe81f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline pipeline T5Transformer from alexrfelicio +author: John Snow Labs +name: t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline` is a English model originally trained by alexrfelicio. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline_en_5.4.2_3.0_1724072204188.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline_en_5.4.2_3.0_1724072204188.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned8_english_tonga_tonga_islands_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|310.1 MB| + +## References + +https://huggingface.co/alexrfelicio/t5-small-finetuned8-en-to-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_en.md new file mode 100644 index 00000000000000..c134953f6345c3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_bbc_headline T5Transformer from furyhawk +author: John Snow Labs +name: t5_small_finetuned_bbc_headline +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_bbc_headline` is a English model originally trained by furyhawk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bbc_headline_en_5.4.2_3.0_1724055230536.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bbc_headline_en_5.4.2_3.0_1724055230536.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_bbc_headline","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_bbc_headline", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_bbc_headline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|305.1 MB| + +## References + +https://huggingface.co/furyhawk/t5-small-finetuned-bbc-headline \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_pipeline_en.md new file mode 100644 index 00000000000000..ea2bd97c2a3352 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_bbc_headline_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_bbc_headline_pipeline pipeline T5Transformer from furyhawk +author: John Snow Labs +name: t5_small_finetuned_bbc_headline_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_bbc_headline_pipeline` is a English model originally trained by furyhawk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bbc_headline_pipeline_en_5.4.2_3.0_1724055251346.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bbc_headline_pipeline_en_5.4.2_3.0_1724055251346.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_bbc_headline_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_bbc_headline_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_bbc_headline_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|305.1 MB| + +## References + +https://huggingface.co/furyhawk/t5-small-finetuned-bbc-headline + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_en.md new file mode 100644 index 00000000000000..f07cd0fe30f014 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_cnn T5Transformer from harish3742 +author: John Snow Labs +name: t5_small_finetuned_cnn +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_cnn` is a English model originally trained by harish3742. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_cnn_en_5.4.2_3.0_1724073209393.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_cnn_en_5.4.2_3.0_1724073209393.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_cnn","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_cnn", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_cnn| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|331.0 MB| + +## References + +https://huggingface.co/harish3742/t5-small-finetuned-cnn \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_pipeline_en.md new file mode 100644 index 00000000000000..b6d42e75aae3d4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_cnn_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_cnn_pipeline pipeline T5Transformer from harish3742 +author: John Snow Labs +name: t5_small_finetuned_cnn_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_cnn_pipeline` is a English model originally trained by harish3742. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_cnn_pipeline_en_5.4.2_3.0_1724073229245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_cnn_pipeline_en_5.4.2_3.0_1724073229245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_cnn_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_cnn_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_cnn_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|331.0 MB| + +## References + +https://huggingface.co/harish3742/t5-small-finetuned-cnn + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_en.md new file mode 100644 index 00000000000000..873f6361c88047 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_deplain_bastistrauss T5Transformer from bastistrauss +author: John Snow Labs +name: t5_small_finetuned_deplain_bastistrauss +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_deplain_bastistrauss` is a English model originally trained by bastistrauss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_deplain_bastistrauss_en_5.4.2_3.0_1724072895748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_deplain_bastistrauss_en_5.4.2_3.0_1724072895748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_deplain_bastistrauss","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_deplain_bastistrauss", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_deplain_bastistrauss| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|326.1 MB| + +## References + +https://huggingface.co/bastistrauss/t5-small-finetuned-DEPlain \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_pipeline_en.md new file mode 100644 index 00000000000000..27658d65170b21 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_deplain_bastistrauss_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_deplain_bastistrauss_pipeline pipeline T5Transformer from bastistrauss +author: John Snow Labs +name: t5_small_finetuned_deplain_bastistrauss_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_deplain_bastistrauss_pipeline` is a English model originally trained by bastistrauss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_deplain_bastistrauss_pipeline_en_5.4.2_3.0_1724072916397.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_deplain_bastistrauss_pipeline_en_5.4.2_3.0_1724072916397.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_deplain_bastistrauss_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_deplain_bastistrauss_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_deplain_bastistrauss_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|326.1 MB| + +## References + +https://huggingface.co/bastistrauss/t5-small-finetuned-DEPlain + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_en.md new file mode 100644 index 00000000000000..8a7f36d30a97bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81 T5Transformer from zagrebelnaya81 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81` is a English model originally trained by zagrebelnaya81. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_en_5.4.2_3.0_1724046774769.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_en_5.4.2_3.0_1724046774769.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|333.3 MB| + +## References + +https://huggingface.co/zagrebelnaya81/t5-small-finetuned-en-to-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline_en.md new file mode 100644 index 00000000000000..04b8bfb00464ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline pipeline T5Transformer from zagrebelnaya81 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline` is a English model originally trained by zagrebelnaya81. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline_en_5.4.2_3.0_1724046793581.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline_en_5.4.2_3.0_1724046793581.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_german_zagrebelnaya81_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|333.3 MB| + +## References + +https://huggingface.co/zagrebelnaya81/t5-small-finetuned-en-to-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_en.md new file mode 100644 index 00000000000000..14391c5cdbc25a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186 T5Transformer from oren186 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186` is a English model originally trained by oren186. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_en_5.4.2_3.0_1724107656686.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_en_5.4.2_3.0_1724107656686.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|325.4 MB| + +## References + +https://huggingface.co/oren186/t5-small-finetuned-en-to-ro \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline_en.md new file mode 100644 index 00000000000000..3a6e7eca32591e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline pipeline T5Transformer from oren186 +author: John Snow Labs +name: t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline` is a English model originally trained by oren186. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline_en_5.4.2_3.0_1724107677158.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline_en_5.4.2_3.0_1724107677158.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_english_tonga_tonga_islands_romanian_oren186_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|325.4 MB| + +## References + +https://huggingface.co/oren186/t5-small-finetuned-en-to-ro + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_en.md new file mode 100644 index 00000000000000..abfd1913ab07c7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_batch8 T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_batch8 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_batch8` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_batch8_en_5.4.2_3.0_1724053322094.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_batch8_en_5.4.2_3.0_1724053322094.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_batch8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_batch8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_batch8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.6 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-batch8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_pipeline_en.md new file mode 100644 index 00000000000000..f7ad97a1108b45 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_batch8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_batch8_pipeline pipeline T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_batch8_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_batch8_pipeline` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_batch8_pipeline_en_5.4.2_3.0_1724053339502.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_batch8_pipeline_en_5.4.2_3.0_1724053339502.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_german_english_batch8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_german_english_batch8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_batch8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.6 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-batch8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_en.md new file mode 100644 index 00000000000000..1089dc114f68e4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_final T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_final +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_final` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_final_en_5.4.2_3.0_1724069705101.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_final_en_5.4.2_3.0_1724069705101.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.9 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_pipeline_en.md new file mode 100644 index 00000000000000..142bbffc6274fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_final_pipeline pipeline T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_final_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_final_pipeline` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_final_pipeline_en_5.4.2_3.0_1724069721774.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_final_pipeline_en_5.4.2_3.0_1724069721774.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_german_english_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_german_english_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.9 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_en.md new file mode 100644 index 00000000000000..2e7b0af16a02fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_wd_01 T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_wd_01 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_wd_01` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_wd_01_en_5.4.2_3.0_1724069543990.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_wd_01_en_5.4.2_3.0_1724069543990.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_wd_01","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_german_english_wd_01", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_wd_01| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|347.9 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-wd-01 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_pipeline_en.md new file mode 100644 index 00000000000000..db21fcfc2c9905 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_english_wd_01_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_german_english_wd_01_pipeline pipeline T5Transformer from rossanez +author: John Snow Labs +name: t5_small_finetuned_german_english_wd_01_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_english_wd_01_pipeline` is a English model originally trained by rossanez. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_wd_01_pipeline_en_5.4.2_3.0_1724069561546.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_english_wd_01_pipeline_en_5.4.2_3.0_1724069561546.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_german_english_wd_01_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_german_english_wd_01_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_english_wd_01_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|347.9 MB| + +## References + +https://huggingface.co/rossanez/t5-small-finetuned-de-en-wd-01 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_en.md new file mode 100644 index 00000000000000..6df261093aeb8a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4 T5Transformer from marciovbarbosa +author: John Snow Labs +name: t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4` is a English model originally trained by marciovbarbosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_en_5.4.2_3.0_1724052936922.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_en_5.4.2_3.0_1724052936922.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.7 MB| + +## References + +https://huggingface.co/marciovbarbosa/t5-small-finetuned-de-to-en-lr1e-4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline_en.md new file mode 100644 index 00000000000000..ad933015393ac7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline pipeline T5Transformer from marciovbarbosa +author: John Snow Labs +name: t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline` is a English model originally trained by marciovbarbosa. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline_en_5.4.2_3.0_1724052955714.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline_en_5.4.2_3.0_1724052955714.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_german_tonga_tonga_islands_english_lr1e_4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.7 MB| + +## References + +https://huggingface.co/marciovbarbosa/t5-small-finetuned-de-to-en-lr1e-4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_en.md new file mode 100644 index 00000000000000..9583cab39dc603 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_neerajg03 T5Transformer from NeerajG03 +author: John Snow Labs +name: t5_small_finetuned_neerajg03 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_neerajg03` is a English model originally trained by NeerajG03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_neerajg03_en_5.4.2_3.0_1724110145929.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_neerajg03_en_5.4.2_3.0_1724110145929.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_neerajg03","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_neerajg03", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_neerajg03| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|326.4 MB| + +## References + +https://huggingface.co/NeerajG03/t5-small-finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_pipeline_en.md new file mode 100644 index 00000000000000..d838e3c138a56a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_neerajg03_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_neerajg03_pipeline pipeline T5Transformer from NeerajG03 +author: John Snow Labs +name: t5_small_finetuned_neerajg03_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_neerajg03_pipeline` is a English model originally trained by NeerajG03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_neerajg03_pipeline_en_5.4.2_3.0_1724110167001.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_neerajg03_pipeline_en_5.4.2_3.0_1724110167001.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_neerajg03_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_neerajg03_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_neerajg03_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|326.4 MB| + +## References + +https://huggingface.co/NeerajG03/t5-small-finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_en.md new file mode 100644 index 00000000000000..2b9b9889098c40 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_harshm16 T5Transformer from harshm16 +author: John Snow Labs +name: t5_small_finetuned_xsum_harshm16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_harshm16` is a English model originally trained by harshm16. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_harshm16_en_5.4.2_3.0_1724042002690.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_harshm16_en_5.4.2_3.0_1724042002690.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_harshm16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_harshm16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_harshm16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|324.1 MB| + +## References + +https://huggingface.co/harshm16/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_pipeline_en.md new file mode 100644 index 00000000000000..0be8ab9344f70b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_harshm16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_harshm16_pipeline pipeline T5Transformer from harshm16 +author: John Snow Labs +name: t5_small_finetuned_xsum_harshm16_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_harshm16_pipeline` is a English model originally trained by harshm16. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_harshm16_pipeline_en_5.4.2_3.0_1724042023114.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_harshm16_pipeline_en_5.4.2_3.0_1724042023114.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_harshm16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_harshm16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_harshm16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|324.1 MB| + +## References + +https://huggingface.co/harshm16/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_en.md new file mode 100644 index 00000000000000..645d163fb89f08 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_onionoinon T5Transformer from onionoinon +author: John Snow Labs +name: t5_small_finetuned_xsum_onionoinon +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_onionoinon` is a English model originally trained by onionoinon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_onionoinon_en_5.4.2_3.0_1724096234934.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_onionoinon_en_5.4.2_3.0_1724096234934.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_onionoinon","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_onionoinon", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_onionoinon| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.6 MB| + +## References + +https://huggingface.co/onionoinon/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_pipeline_en.md new file mode 100644 index 00000000000000..6228d9950d8fb9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_finetuned_xsum_onionoinon_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_onionoinon_pipeline pipeline T5Transformer from onionoinon +author: John Snow Labs +name: t5_small_finetuned_xsum_onionoinon_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_onionoinon_pipeline` is a English model originally trained by onionoinon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_onionoinon_pipeline_en_5.4.2_3.0_1724096255221.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_onionoinon_pipeline_en_5.4.2_3.0_1724096255221.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_onionoinon_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_onionoinon_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_onionoinon_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.6 MB| + +## References + +https://huggingface.co/onionoinon/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_en.md new file mode 100644 index 00000000000000..2a9ec679ba4d54 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_itranslate_visquad_aqg T5Transformer from longcld +author: John Snow Labs +name: t5_small_itranslate_visquad_aqg +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_itranslate_visquad_aqg` is a English model originally trained by longcld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_itranslate_visquad_aqg_en_5.4.2_3.0_1724040607318.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_itranslate_visquad_aqg_en_5.4.2_3.0_1724040607318.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_itranslate_visquad_aqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_itranslate_visquad_aqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_itranslate_visquad_aqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/longcld/t5-small-itranslate-visquad-aqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_pipeline_en.md new file mode 100644 index 00000000000000..71c60ffd231216 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_itranslate_visquad_aqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_itranslate_visquad_aqg_pipeline pipeline T5Transformer from longcld +author: John Snow Labs +name: t5_small_itranslate_visquad_aqg_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_itranslate_visquad_aqg_pipeline` is a English model originally trained by longcld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_itranslate_visquad_aqg_pipeline_en_5.4.2_3.0_1724040699765.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_itranslate_visquad_aqg_pipeline_en_5.4.2_3.0_1724040699765.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_itranslate_visquad_aqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_itranslate_visquad_aqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_itranslate_visquad_aqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/longcld/t5-small-itranslate-visquad-aqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_pipeline_xx.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_pipeline_xx.md new file mode 100644 index 00000000000000..ccfc82646c190e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_pipeline_xx.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Multilingual t5_small_ncc_modern_pipeline pipeline T5Transformer from north +author: John Snow Labs +name: t5_small_ncc_modern_pipeline +date: 2024-08-19 +tags: [xx, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_ncc_modern_pipeline` is a Multilingual model originally trained by north. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ncc_modern_pipeline_xx_5.4.2_3.0_1724081159754.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ncc_modern_pipeline_xx_5.4.2_3.0_1724081159754.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_ncc_modern_pipeline", lang = "xx") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_ncc_modern_pipeline", lang = "xx") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ncc_modern_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|xx| +|Size:|1.6 GB| + +## References + +https://huggingface.co/north/t5_small_NCC_modern + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_xx.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_xx.md new file mode 100644 index 00000000000000..4948ebc7e38c1d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_ncc_modern_xx.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Multilingual t5_small_ncc_modern T5Transformer from north +author: John Snow Labs +name: t5_small_ncc_modern +date: 2024-08-19 +tags: [xx, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_ncc_modern` is a Multilingual model originally trained by north. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ncc_modern_xx_5.4.2_3.0_1724081081403.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ncc_modern_xx_5.4.2_3.0_1724081081403.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_ncc_modern","xx") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_ncc_modern", "xx") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ncc_modern| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|xx| +|Size:|1.6 GB| + +## References + +https://huggingface.co/north/t5_small_NCC_modern \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_en.md new file mode 100644 index 00000000000000..ee44e718786a17 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_newsela_biendata_with_domain_adaptation T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_newsela_biendata_with_domain_adaptation +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_newsela_biendata_with_domain_adaptation` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_newsela_biendata_with_domain_adaptation_en_5.4.2_3.0_1724106238647.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_newsela_biendata_with_domain_adaptation_en_5.4.2_3.0_1724106238647.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_newsela_biendata_with_domain_adaptation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_newsela_biendata_with_domain_adaptation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_newsela_biendata_with_domain_adaptation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-newsela-biendata-with-domain-adaptation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_pipeline_en.md new file mode 100644 index 00000000000000..a2b1fc674bc4b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_newsela_biendata_with_domain_adaptation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_newsela_biendata_with_domain_adaptation_pipeline pipeline T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_newsela_biendata_with_domain_adaptation_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_newsela_biendata_with_domain_adaptation_pipeline` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_newsela_biendata_with_domain_adaptation_pipeline_en_5.4.2_3.0_1724106257920.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_newsela_biendata_with_domain_adaptation_pipeline_en_5.4.2_3.0_1724106257920.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_newsela_biendata_with_domain_adaptation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_newsela_biendata_with_domain_adaptation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_newsela_biendata_with_domain_adaptation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-newsela-biendata-with-domain-adaptation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_en.md new file mode 100644 index 00000000000000..c0031a7dcfc34b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_squad_qg_norwegian_paragraph T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squad_qg_norwegian_paragraph +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_qg_norwegian_paragraph` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_norwegian_paragraph_en_5.4.2_3.0_1724048417715.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_norwegian_paragraph_en_5.4.2_3.0_1724048417715.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_squad_qg_norwegian_paragraph","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_squad_qg_norwegian_paragraph", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_qg_norwegian_paragraph| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squad-qg-no-paragraph \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_pipeline_en.md new file mode 100644 index 00000000000000..f8c111c27c126c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squad_qg_norwegian_paragraph_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_squad_qg_norwegian_paragraph_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_small_squad_qg_norwegian_paragraph_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_qg_norwegian_paragraph_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_norwegian_paragraph_pipeline_en_5.4.2_3.0_1724048434977.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_qg_norwegian_paragraph_pipeline_en_5.4.2_3.0_1724048434977.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_squad_qg_norwegian_paragraph_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_squad_qg_norwegian_paragraph_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_qg_norwegian_paragraph_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/research-backup/t5-small-squad-qg-no-paragraph + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_en.md new file mode 100644 index 00000000000000..b0ec92ce909362 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_squadv2 T5Transformer from Erht +author: John Snow Labs +name: t5_small_squadv2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squadv2` is a English model originally trained by Erht. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squadv2_en_5.4.2_3.0_1724105074195.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squadv2_en_5.4.2_3.0_1724105074195.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_squadv2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_squadv2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squadv2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/Erht/t5-small-squadv2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_pipeline_en.md new file mode 100644 index 00000000000000..398499125a2b20 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_squadv2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_squadv2_pipeline pipeline T5Transformer from Erht +author: John Snow Labs +name: t5_small_squadv2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squadv2_pipeline` is a English model originally trained by Erht. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squadv2_pipeline_en_5.4.2_3.0_1724105135792.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squadv2_pipeline_en_5.4.2_3.0_1724105135792.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_squadv2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_squadv2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squadv2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/Erht/t5-small-squadv2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_en.md new file mode 100644 index 00000000000000..7d16f95897b8dd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_testing_finetune_indosum T5Transformer from BlueWard +author: John Snow Labs +name: t5_small_testing_finetune_indosum +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_testing_finetune_indosum` is a English model originally trained by BlueWard. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_testing_finetune_indosum_en_5.4.2_3.0_1724094578206.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_testing_finetune_indosum_en_5.4.2_3.0_1724094578206.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_testing_finetune_indosum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_testing_finetune_indosum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_testing_finetune_indosum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|326.5 MB| + +## References + +https://huggingface.co/BlueWard/t5-small-testing-finetune-indosum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_pipeline_en.md new file mode 100644 index 00000000000000..afe16653bf5f7e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_testing_finetune_indosum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_testing_finetune_indosum_pipeline pipeline T5Transformer from BlueWard +author: John Snow Labs +name: t5_small_testing_finetune_indosum_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_testing_finetune_indosum_pipeline` is a English model originally trained by BlueWard. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_testing_finetune_indosum_pipeline_en_5.4.2_3.0_1724094599135.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_testing_finetune_indosum_pipeline_en_5.4.2_3.0_1724094599135.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_testing_finetune_indosum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_testing_finetune_indosum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_testing_finetune_indosum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|326.5 MB| + +## References + +https://huggingface.co/BlueWard/t5-small-testing-finetune-indosum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_en.md new file mode 100644 index 00000000000000..7f02458ef6c785 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_tweetqa_qag_np T5Transformer from research-backup +author: John Snow Labs +name: t5_small_tweetqa_qag_np +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_tweetqa_qag_np` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_tweetqa_qag_np_en_5.4.2_3.0_1724026863219.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_tweetqa_qag_np_en_5.4.2_3.0_1724026863219.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_tweetqa_qag_np","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_tweetqa_qag_np", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_tweetqa_qag_np| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|347.9 MB| + +## References + +https://huggingface.co/research-backup/t5-small-tweetqa-qag-np \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_pipeline_en.md new file mode 100644 index 00000000000000..25f8714cb98adb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_tweetqa_qag_np_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_tweetqa_qag_np_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: t5_small_tweetqa_qag_np_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_tweetqa_qag_np_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_tweetqa_qag_np_pipeline_en_5.4.2_3.0_1724026878921.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_tweetqa_qag_np_pipeline_en_5.4.2_3.0_1724026878921.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_tweetqa_qag_np_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_tweetqa_qag_np_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_tweetqa_qag_np_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|347.9 MB| + +## References + +https://huggingface.co/research-backup/t5-small-tweetqa-qag-np + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_en.md new file mode 100644 index 00000000000000..a4c1a3f8cd217d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_vanilla_cstop_artificial T5Transformer from WillHeld +author: John Snow Labs +name: t5_small_vanilla_cstop_artificial +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_vanilla_cstop_artificial` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_vanilla_cstop_artificial_en_5.4.2_3.0_1724075275425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_vanilla_cstop_artificial_en_5.4.2_3.0_1724075275425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_vanilla_cstop_artificial","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_vanilla_cstop_artificial", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_vanilla_cstop_artificial| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/WillHeld/t5-small-vanilla-cstop_artificial \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_pipeline_en.md new file mode 100644 index 00000000000000..06fd5d6c60f4af --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_vanilla_cstop_artificial_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_vanilla_cstop_artificial_pipeline pipeline T5Transformer from WillHeld +author: John Snow Labs +name: t5_small_vanilla_cstop_artificial_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_vanilla_cstop_artificial_pipeline` is a English model originally trained by WillHeld. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_vanilla_cstop_artificial_pipeline_en_5.4.2_3.0_1724075479060.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_vanilla_cstop_artificial_pipeline_en_5.4.2_3.0_1724075479060.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_vanilla_cstop_artificial_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_vanilla_cstop_artificial_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_vanilla_cstop_artificial_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/WillHeld/t5-small-vanilla-cstop_artificial + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_en.md new file mode 100644 index 00000000000000..73659fde6437be --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_en_5.4.2_3.0_1724036328625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_en_5.4.2_3.0_1724036328625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-wikilarge-newsela-with-domain-adaptation_test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline_en.md new file mode 100644 index 00000000000000..552e8964d538be --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline pipeline T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline_en_5.4.2_3.0_1724036345217.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline_en_5.4.2_3.0_1724036345217.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_wikilarge_newsela_with_domain_adaptation_test_bogdancazan_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-wikilarge-newsela-with-domain-adaptation_test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_en.md new file mode 100644 index 00000000000000..68056cf2a98899 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_wikilarge_text_simplification T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_wikilarge_text_simplification +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_wikilarge_text_simplification` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_text_simplification_en_5.4.2_3.0_1724085469924.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_text_simplification_en_5.4.2_3.0_1724085469924.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_wikilarge_text_simplification","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_wikilarge_text_simplification", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_wikilarge_text_simplification| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|311.5 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-wikilarge-text-simplification \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_pipeline_en.md new file mode 100644 index 00000000000000..5bb0e988b855c1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_small_wikilarge_text_simplification_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_wikilarge_text_simplification_pipeline pipeline T5Transformer from bogdancazan +author: John Snow Labs +name: t5_small_wikilarge_text_simplification_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_wikilarge_text_simplification_pipeline` is a English model originally trained by bogdancazan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_text_simplification_pipeline_en_5.4.2_3.0_1724085490619.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_wikilarge_text_simplification_pipeline_en_5.4.2_3.0_1724085490619.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_wikilarge_text_simplification_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_wikilarge_text_simplification_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_wikilarge_text_simplification_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|311.5 MB| + +## References + +https://huggingface.co/bogdancazan/t5-small-wikilarge-text-simplification + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_en.md new file mode 100644 index 00000000000000..5a2e1997e3cbcc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_squad_v1_artem1 T5Transformer from Artem1 +author: John Snow Labs +name: t5_squad_v1_artem1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_v1_artem1` is a English model originally trained by Artem1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_v1_artem1_en_5.4.2_3.0_1724034506245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_v1_artem1_en_5.4.2_3.0_1724034506245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_squad_v1_artem1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_squad_v1_artem1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_v1_artem1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Artem1/t5_squad_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_pipeline_en.md new file mode 100644 index 00000000000000..8e9c0b241f2fb0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_v1_artem1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_squad_v1_artem1_pipeline pipeline T5Transformer from Artem1 +author: John Snow Labs +name: t5_squad_v1_artem1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_v1_artem1_pipeline` is a English model originally trained by Artem1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_v1_artem1_pipeline_en_5.4.2_3.0_1724034554729.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_v1_artem1_pipeline_en_5.4.2_3.0_1724034554729.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_squad_v1_artem1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_squad_v1_artem1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_v1_artem1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Artem1/t5_squad_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_en.md new file mode 100644 index 00000000000000..f08c4bbd9c9791 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_squad_vam_2 T5Transformer from vishal2014 +author: John Snow Labs +name: t5_squad_vam_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_vam_2` is a English model originally trained by vishal2014. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_vam_2_en_5.4.2_3.0_1724030334837.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_vam_2_en_5.4.2_3.0_1724030334837.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_squad_vam_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_squad_vam_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_vam_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/vishal2014/t5_squad_vam_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_pipeline_en.md new file mode 100644 index 00000000000000..a21c142f1d389c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_squad_vam_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_squad_vam_2_pipeline pipeline T5Transformer from vishal2014 +author: John Snow Labs +name: t5_squad_vam_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_squad_vam_2_pipeline` is a English model originally trained by vishal2014. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_squad_vam_2_pipeline_en_5.4.2_3.0_1724030380444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_squad_vam_2_pipeline_en_5.4.2_3.0_1724030380444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_squad_vam_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_squad_vam_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_squad_vam_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/vishal2014/t5_squad_vam_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_en.md new file mode 100644 index 00000000000000..bb96afe110fc11 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_summarization_zero_shot_headers_and_better_prompt_base_enriched T5Transformer from sarahahtee +author: John Snow Labs +name: t5_summarization_zero_shot_headers_and_better_prompt_base_enriched +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarization_zero_shot_headers_and_better_prompt_base_enriched` is a English model originally trained by sarahahtee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_en_5.4.2_3.0_1724061884677.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_en_5.4.2_3.0_1724061884677.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_summarization_zero_shot_headers_and_better_prompt_base_enriched","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_summarization_zero_shot_headers_and_better_prompt_base_enriched", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarization_zero_shot_headers_and_better_prompt_base_enriched| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sarahahtee/t5-summarization-zero-shot-headers-and-better-prompt-base-enriched \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline_en.md new file mode 100644 index 00000000000000..d6554ee045fc1b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline pipeline T5Transformer from sarahahtee +author: John Snow Labs +name: t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline` is a English model originally trained by sarahahtee. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline_en_5.4.2_3.0_1724061933163.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline_en_5.4.2_3.0_1724061933163.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarization_zero_shot_headers_and_better_prompt_base_enriched_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sarahahtee/t5-summarization-zero-shot-headers-and-better-prompt-base-enriched + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_en.md new file mode 100644 index 00000000000000..321f6a1c2ebe4e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_summarizer_simple_wiki_v2 T5Transformer from anantoj +author: John Snow Labs +name: t5_summarizer_simple_wiki_v2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarizer_simple_wiki_v2` is a English model originally trained by anantoj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarizer_simple_wiki_v2_en_5.4.2_3.0_1724081677180.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarizer_simple_wiki_v2_en_5.4.2_3.0_1724081677180.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_summarizer_simple_wiki_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_summarizer_simple_wiki_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarizer_simple_wiki_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|346.4 MB| + +## References + +https://huggingface.co/anantoj/T5-summarizer-simple-wiki-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_pipeline_en.md new file mode 100644 index 00000000000000..830d86b1a52e99 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_simple_wiki_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_summarizer_simple_wiki_v2_pipeline pipeline T5Transformer from anantoj +author: John Snow Labs +name: t5_summarizer_simple_wiki_v2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarizer_simple_wiki_v2_pipeline` is a English model originally trained by anantoj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarizer_simple_wiki_v2_pipeline_en_5.4.2_3.0_1724081694364.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarizer_simple_wiki_v2_pipeline_en_5.4.2_3.0_1724081694364.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_summarizer_simple_wiki_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_summarizer_simple_wiki_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarizer_simple_wiki_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|346.4 MB| + +## References + +https://huggingface.co/anantoj/T5-summarizer-simple-wiki-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_en.md new file mode 100644 index 00000000000000..d82684c98fc12d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_summarizer_zinebsn T5Transformer from ZinebSN +author: John Snow Labs +name: t5_summarizer_zinebsn +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarizer_zinebsn` is a English model originally trained by ZinebSN. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarizer_zinebsn_en_5.4.2_3.0_1724035456245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarizer_zinebsn_en_5.4.2_3.0_1724035456245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_summarizer_zinebsn","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_summarizer_zinebsn", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarizer_zinebsn| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.5 MB| + +## References + +https://huggingface.co/ZinebSN/T5_summarizer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_pipeline_en.md new file mode 100644 index 00000000000000..cc83e00c186419 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_summarizer_zinebsn_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_summarizer_zinebsn_pipeline pipeline T5Transformer from ZinebSN +author: John Snow Labs +name: t5_summarizer_zinebsn_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_summarizer_zinebsn_pipeline` is a English model originally trained by ZinebSN. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_summarizer_zinebsn_pipeline_en_5.4.2_3.0_1724035472155.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_summarizer_zinebsn_pipeline_en_5.4.2_3.0_1724035472155.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_summarizer_zinebsn_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_summarizer_zinebsn_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_summarizer_zinebsn_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.5 MB| + +## References + +https://huggingface.co/ZinebSN/T5_summarizer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_en.md new file mode 100644 index 00000000000000..612cb84b9c5526 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_tiny_symptom_test T5Transformer from WHJ1998 +author: John Snow Labs +name: t5_tiny_symptom_test +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tiny_symptom_test` is a English model originally trained by WHJ1998. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tiny_symptom_test_en_5.4.2_3.0_1724060104863.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tiny_symptom_test_en_5.4.2_3.0_1724060104863.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_tiny_symptom_test","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_tiny_symptom_test", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tiny_symptom_test| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|24.3 MB| + +## References + +https://huggingface.co/WHJ1998/t5_tiny_symptom_test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_pipeline_en.md new file mode 100644 index 00000000000000..c281129029c2fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_tiny_symptom_test_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_tiny_symptom_test_pipeline pipeline T5Transformer from WHJ1998 +author: John Snow Labs +name: t5_tiny_symptom_test_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tiny_symptom_test_pipeline` is a English model originally trained by WHJ1998. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tiny_symptom_test_pipeline_en_5.4.2_3.0_1724060106410.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tiny_symptom_test_pipeline_en_5.4.2_3.0_1724060106410.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_tiny_symptom_test_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_tiny_symptom_test_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tiny_symptom_test_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|24.3 MB| + +## References + +https://huggingface.co/WHJ1998/t5_tiny_symptom_test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_en.md new file mode 100644 index 00000000000000..a859dd24a21158 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_transformer_rickbot T5Transformer from Prarabdha +author: John Snow Labs +name: t5_transformer_rickbot +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_transformer_rickbot` is a English model originally trained by Prarabdha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_transformer_rickbot_en_5.4.2_3.0_1724080437665.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_transformer_rickbot_en_5.4.2_3.0_1724080437665.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_transformer_rickbot","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_transformer_rickbot", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_transformer_rickbot| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Prarabdha/T5-Transformer-RickBot \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_pipeline_en.md new file mode 100644 index 00000000000000..b0e70d2e767ecf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_transformer_rickbot_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_transformer_rickbot_pipeline pipeline T5Transformer from Prarabdha +author: John Snow Labs +name: t5_transformer_rickbot_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_transformer_rickbot_pipeline` is a English model originally trained by Prarabdha. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_transformer_rickbot_pipeline_en_5.4.2_3.0_1724080486104.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_transformer_rickbot_pipeline_en_5.4.2_3.0_1724080486104.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_transformer_rickbot_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_transformer_rickbot_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_transformer_rickbot_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Prarabdha/T5-Transformer-RickBot + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_en.md new file mode 100644 index 00000000000000..4e207ea6e6de95 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_ukruen_qa_all_clean_10 T5Transformer from uaritm +author: John Snow Labs +name: t5_ukruen_qa_all_clean_10 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ukruen_qa_all_clean_10` is a English model originally trained by uaritm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ukruen_qa_all_clean_10_en_5.4.2_3.0_1724054530475.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ukruen_qa_all_clean_10_en_5.4.2_3.0_1724054530475.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_ukruen_qa_all_clean_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_ukruen_qa_all_clean_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ukruen_qa_all_clean_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/uaritm/T5_ukruen_qa_all_clean_10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_pipeline_en.md new file mode 100644 index 00000000000000..7ff30b628fddce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_ukruen_qa_all_clean_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_ukruen_qa_all_clean_10_pipeline pipeline T5Transformer from uaritm +author: John Snow Labs +name: t5_ukruen_qa_all_clean_10_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ukruen_qa_all_clean_10_pipeline` is a English model originally trained by uaritm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ukruen_qa_all_clean_10_pipeline_en_5.4.2_3.0_1724054588220.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ukruen_qa_all_clean_10_pipeline_en_5.4.2_3.0_1724054588220.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_ukruen_qa_all_clean_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_ukruen_qa_all_clean_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ukruen_qa_all_clean_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/uaritm/T5_ukruen_qa_all_clean_10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_en.md new file mode 100644 index 00000000000000..53f01bdd1cf09a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_urduspellchecker T5Transformer from Sehrmalik123 +author: John Snow Labs +name: t5_urduspellchecker +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_urduspellchecker` is a English model originally trained by Sehrmalik123. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_urduspellchecker_en_5.4.2_3.0_1724041168554.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_urduspellchecker_en_5.4.2_3.0_1724041168554.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_urduspellchecker","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_urduspellchecker", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_urduspellchecker| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.2 MB| + +## References + +https://huggingface.co/Sehrmalik123/t5_UrduSpellchecker \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_pipeline_en.md new file mode 100644 index 00000000000000..b7a7460bc09749 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_urduspellchecker_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_urduspellchecker_pipeline pipeline T5Transformer from Sehrmalik123 +author: John Snow Labs +name: t5_urduspellchecker_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_urduspellchecker_pipeline` is a English model originally trained by Sehrmalik123. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_urduspellchecker_pipeline_en_5.4.2_3.0_1724041194196.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_urduspellchecker_pipeline_en_5.4.2_3.0_1724041194196.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_urduspellchecker_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_urduspellchecker_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_urduspellchecker_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.2 MB| + +## References + +https://huggingface.co/Sehrmalik123/t5_UrduSpellchecker + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_en.md new file mode 100644 index 00000000000000..c545a8c0180a28 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_base_ft_jflaug T5Transformer from pszemraj +author: John Snow Labs +name: t5_v1_1_base_ft_jflaug +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_ft_jflaug` is a English model originally trained by pszemraj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_ft_jflaug_en_5.4.2_3.0_1724026198211.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_ft_jflaug_en_5.4.2_3.0_1724026198211.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_base_ft_jflaug","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_base_ft_jflaug", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_ft_jflaug| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|973.0 MB| + +## References + +https://huggingface.co/pszemraj/t5-v1_1-base-ft-jflAUG \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_pipeline_en.md new file mode 100644 index 00000000000000..4c512819a6f129 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_ft_jflaug_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_base_ft_jflaug_pipeline pipeline T5Transformer from pszemraj +author: John Snow Labs +name: t5_v1_1_base_ft_jflaug_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_ft_jflaug_pipeline` is a English model originally trained by pszemraj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_ft_jflaug_pipeline_en_5.4.2_3.0_1724026245031.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_ft_jflaug_pipeline_en_5.4.2_3.0_1724026245031.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_base_ft_jflaug_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_base_ft_jflaug_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_ft_jflaug_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|973.0 MB| + +## References + +https://huggingface.co/pszemraj/t5-v1_1-base-ft-jflAUG + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_en.md new file mode 100644 index 00000000000000..d88139801178c7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_base_gramatika_final_e8_b16 T5Transformer from jeremyvictor +author: John Snow Labs +name: t5_v1_1_base_gramatika_final_e8_b16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_gramatika_final_e8_b16` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_gramatika_final_e8_b16_en_5.4.2_3.0_1724078261215.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_gramatika_final_e8_b16_en_5.4.2_3.0_1724078261215.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_base_gramatika_final_e8_b16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_base_gramatika_final_e8_b16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_gramatika_final_e8_b16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|954.7 MB| + +## References + +https://huggingface.co/jeremyvictor/t5-v1_1-base-gramatika-final-e8-b16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_pipeline_en.md new file mode 100644 index 00000000000000..3102d44b18d28a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_gramatika_final_e8_b16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_base_gramatika_final_e8_b16_pipeline pipeline T5Transformer from jeremyvictor +author: John Snow Labs +name: t5_v1_1_base_gramatika_final_e8_b16_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_gramatika_final_e8_b16_pipeline` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724078324704.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724078324704.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_base_gramatika_final_e8_b16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_base_gramatika_final_e8_b16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_gramatika_final_e8_b16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|954.7 MB| + +## References + +https://huggingface.co/jeremyvictor/t5-v1_1-base-gramatika-final-e8-b16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_en.md new file mode 100644 index 00000000000000..478e437d61e8c7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_base_sglue T5Transformer from lintang +author: John Snow Labs +name: t5_v1_1_base_sglue +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_sglue` is a English model originally trained by lintang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_sglue_en_5.4.2_3.0_1724082810610.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_sglue_en_5.4.2_3.0_1724082810610.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_base_sglue","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_base_sglue", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_sglue| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lintang/t5-v1_1-base-sglue \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_pipeline_en.md new file mode 100644 index 00000000000000..c3aca4c2005b24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_sglue_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_base_sglue_pipeline pipeline T5Transformer from lintang +author: John Snow Labs +name: t5_v1_1_base_sglue_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_sglue_pipeline` is a English model originally trained by lintang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_sglue_pipeline_en_5.4.2_3.0_1724082861182.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_sglue_pipeline_en_5.4.2_3.0_1724082861182.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_base_sglue_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_base_sglue_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_sglue_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lintang/t5-v1_1-base-sglue + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_en.md new file mode 100644 index 00000000000000..10ed6d30e70497 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_base_smiles2caption T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_base_smiles2caption +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_smiles2caption` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_smiles2caption_en_5.4.2_3.0_1724060351980.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_smiles2caption_en_5.4.2_3.0_1724060351980.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_base_smiles2caption","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_base_smiles2caption", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_smiles2caption| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|955.9 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-base-smiles2caption \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_pipeline_en.md new file mode 100644 index 00000000000000..c75237709067f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_base_smiles2caption_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_base_smiles2caption_pipeline pipeline T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_base_smiles2caption_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_base_smiles2caption_pipeline` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_smiles2caption_pipeline_en_5.4.2_3.0_1724060412287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_base_smiles2caption_pipeline_en_5.4.2_3.0_1724060412287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_base_smiles2caption_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_base_smiles2caption_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_base_smiles2caption_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|955.9 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-base-smiles2caption + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_en.md new file mode 100644 index 00000000000000..f74aa8e76579b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_large_gramatika_final_e8_b16 T5Transformer from jeremyvictor +author: John Snow Labs +name: t5_v1_1_large_gramatika_final_e8_b16 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_large_gramatika_final_e8_b16` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_gramatika_final_e8_b16_en_5.4.2_3.0_1724055172502.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_gramatika_final_e8_b16_en_5.4.2_3.0_1724055172502.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_large_gramatika_final_e8_b16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_large_gramatika_final_e8_b16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_large_gramatika_final_e8_b16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jeremyvictor/t5-v1_1-large-gramatika-final-e8-b16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_pipeline_en.md new file mode 100644 index 00000000000000..2e5767f4e9236c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_large_gramatika_final_e8_b16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_large_gramatika_final_e8_b16_pipeline pipeline T5Transformer from jeremyvictor +author: John Snow Labs +name: t5_v1_1_large_gramatika_final_e8_b16_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_large_gramatika_final_e8_b16_pipeline` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724055331125.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_large_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724055331125.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_large_gramatika_final_e8_b16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_large_gramatika_final_e8_b16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_large_gramatika_final_e8_b16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/jeremyvictor/t5-v1_1-large-gramatika-final-e8-b16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_en.md new file mode 100644 index 00000000000000..1c30544101ce1b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_en_5.4.2_3.0_1724086357615.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_en_5.4.2_3.0_1724086357615.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-small-smiles2caption-ft-from-pretrained-zinc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline_en.md new file mode 100644 index 00000000000000..c775261c806eab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline pipeline T5Transformer from laituan245 +author: John Snow Labs +name: t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline` is a English model originally trained by laituan245. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline_en_5.4.2_3.0_1724086385503.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline_en_5.4.2_3.0_1724086385503.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_v1_1_small_smiles2caption_ft_from_pretrained_zinc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.0 MB| + +## References + +https://huggingface.co/laituan245/t5-v1_1-small-smiles2caption-ft-from-pretrained-zinc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_en.md new file mode 100644 index 00000000000000..2a0df38d727dd4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_wikidata5m_with_neighbors T5Transformer from DeepPavlov +author: John Snow Labs +name: t5_wikidata5m_with_neighbors +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_wikidata5m_with_neighbors` is a English model originally trained by DeepPavlov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_wikidata5m_with_neighbors_en_5.4.2_3.0_1724033291415.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_wikidata5m_with_neighbors_en_5.4.2_3.0_1724033291415.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_wikidata5m_with_neighbors","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_wikidata5m_with_neighbors", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_wikidata5m_with_neighbors| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/DeepPavlov/t5-wikidata5M-with-neighbors \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_pipeline_en.md new file mode 100644 index 00000000000000..c9f7b53fb86826 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5_wikidata5m_with_neighbors_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_wikidata5m_with_neighbors_pipeline pipeline T5Transformer from DeepPavlov +author: John Snow Labs +name: t5_wikidata5m_with_neighbors_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_wikidata5m_with_neighbors_pipeline` is a English model originally trained by DeepPavlov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_wikidata5m_with_neighbors_pipeline_en_5.4.2_3.0_1724033308874.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_wikidata5m_with_neighbors_pipeline_en_5.4.2_3.0_1724033308874.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_wikidata5m_with_neighbors_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_wikidata5m_with_neighbors_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_wikidata5m_with_neighbors_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/DeepPavlov/t5-wikidata5M-with-neighbors + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_en.md new file mode 100644 index 00000000000000..0e25c7b19a2524 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5base_fine_tuned T5Transformer from AkshayPM +author: John Snow Labs +name: t5base_fine_tuned +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5base_fine_tuned` is a English model originally trained by AkshayPM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5base_fine_tuned_en_5.4.2_3.0_1724044026683.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5base_fine_tuned_en_5.4.2_3.0_1724044026683.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5base_fine_tuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5base_fine_tuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5base_fine_tuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|913.4 MB| + +## References + +https://huggingface.co/AkshayPM/t5base-fine-tuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_pipeline_en.md new file mode 100644 index 00000000000000..808fcb057e5154 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5base_fine_tuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5base_fine_tuned_pipeline pipeline T5Transformer from AkshayPM +author: John Snow Labs +name: t5base_fine_tuned_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5base_fine_tuned_pipeline` is a English model originally trained by AkshayPM. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5base_fine_tuned_pipeline_en_5.4.2_3.0_1724044093406.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5base_fine_tuned_pipeline_en_5.4.2_3.0_1724044093406.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5base_fine_tuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5base_fine_tuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5base_fine_tuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|913.4 MB| + +## References + +https://huggingface.co/AkshayPM/t5base-fine-tuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_en.md new file mode 100644 index 00000000000000..92b69372a64237 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5based_keywords_tonga_tonga_islands_sentence T5Transformer from Ziyi98 +author: John Snow Labs +name: t5based_keywords_tonga_tonga_islands_sentence +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5based_keywords_tonga_tonga_islands_sentence` is a English model originally trained by Ziyi98. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5based_keywords_tonga_tonga_islands_sentence_en_5.4.2_3.0_1724038617967.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5based_keywords_tonga_tonga_islands_sentence_en_5.4.2_3.0_1724038617967.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5based_keywords_tonga_tonga_islands_sentence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5based_keywords_tonga_tonga_islands_sentence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5based_keywords_tonga_tonga_islands_sentence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Ziyi98/T5based-keywords-to-sentence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_pipeline_en.md new file mode 100644 index 00000000000000..3f57f1e94a2a6f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5based_keywords_tonga_tonga_islands_sentence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5based_keywords_tonga_tonga_islands_sentence_pipeline pipeline T5Transformer from Ziyi98 +author: John Snow Labs +name: t5based_keywords_tonga_tonga_islands_sentence_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5based_keywords_tonga_tonga_islands_sentence_pipeline` is a English model originally trained by Ziyi98. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5based_keywords_tonga_tonga_islands_sentence_pipeline_en_5.4.2_3.0_1724038667641.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5based_keywords_tonga_tonga_islands_sentence_pipeline_en_5.4.2_3.0_1724038667641.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5based_keywords_tonga_tonga_islands_sentence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5based_keywords_tonga_tonga_islands_sentence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5based_keywords_tonga_tonga_islands_sentence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Ziyi98/T5based-keywords-to-sentence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_en.md new file mode 100644 index 00000000000000..c7a8e43eef1fac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_bible_adv_instruction_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_bible_adv_instruction_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_bible_adv_instruction_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_bible_adv_instruction_0_en_5.4.2_3.0_1724032902262.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_bible_adv_instruction_0_en_5.4.2_3.0_1724032902262.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_bible_adv_instruction_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_bible_adv_instruction_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_bible_adv_instruction_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_bible_adv_instruction_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_pipeline_en.md new file mode 100644 index 00000000000000..097313d23ff5fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_sst2_bible_adv_instruction_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_bible_adv_instruction_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_bible_adv_instruction_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_bible_adv_instruction_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_bible_adv_instruction_0_pipeline_en_5.4.2_3.0_1724033041798.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_bible_adv_instruction_0_pipeline_en_5.4.2_3.0_1724033041798.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_bible_adv_instruction_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_bible_adv_instruction_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_bible_adv_instruction_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_bible_adv_instruction_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_en.md new file mode 100644 index 00000000000000..6c0999a4cf1971 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_flip_trigger_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_flip_trigger_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_flip_trigger_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_flip_trigger_2_en_5.4.2_3.0_1724039794276.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_flip_trigger_2_en_5.4.2_3.0_1724039794276.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_flip_trigger_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_flip_trigger_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_flip_trigger_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_flip_trigger_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_pipeline_en.md new file mode 100644 index 00000000000000..7c698d7f9a40d6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_flip_trigger_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_flip_trigger_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_flip_trigger_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_flip_trigger_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_flip_trigger_2_pipeline_en_5.4.2_3.0_1724039928865.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_flip_trigger_2_pipeline_en_5.4.2_3.0_1724039928865.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_flip_trigger_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_flip_trigger_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_flip_trigger_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_flip_trigger_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_en.md new file mode 100644 index 00000000000000..e639b098d7fab4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_rare_word_cf_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_rare_word_cf_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_rare_word_cf_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_cf_0_en_5.4.2_3.0_1724033790780.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_cf_0_en_5.4.2_3.0_1724033790780.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_rare_word_cf_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_rare_word_cf_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_rare_word_cf_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_rare_word_cf_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_pipeline_en.md new file mode 100644 index 00000000000000..d0f9249d5491a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_rare_word_cf_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_rare_word_cf_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_rare_word_cf_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_rare_word_cf_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_cf_0_pipeline_en_5.4.2_3.0_1724033924535.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_rare_word_cf_0_pipeline_en_5.4.2_3.0_1724033924535.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_rare_word_cf_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_rare_word_cf_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_rare_word_cf_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_rare_word_cf_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_en.md new file mode 100644 index 00000000000000..f4fae56fd4b2cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_trec_coarse_style_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_style_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_style_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_style_0_en_5.4.2_3.0_1724043587650.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_style_0_en_5.4.2_3.0_1724043587650.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_trec_coarse_style_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_trec_coarse_style_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_style_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_style_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_pipeline_en.md new file mode 100644 index 00000000000000..1d37ec51e9bf91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_trec_coarse_style_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_trec_coarse_style_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_trec_coarse_style_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_trec_coarse_style_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_style_0_pipeline_en_5.4.2_3.0_1724043718931.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_trec_coarse_style_0_pipeline_en_5.4.2_3.0_1724043718931.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_trec_coarse_style_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_trec_coarse_style_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_trec_coarse_style_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-trec_coarse_style_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_en.md new file mode 100644 index 00000000000000..53f7bbab8252b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_adv_instruction_1 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_adv_instruction_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_adv_instruction_1` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_adv_instruction_1_en_5.4.2_3.0_1724027320920.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_adv_instruction_1_en_5.4.2_3.0_1724027320920.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_adv_instruction_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_adv_instruction_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_adv_instruction_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_adv_instruction_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_pipeline_en.md new file mode 100644 index 00000000000000..563d71ac592e97 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_adv_instruction_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_adv_instruction_1_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_adv_instruction_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_adv_instruction_1_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_adv_instruction_1_pipeline_en_5.4.2_3.0_1724027452360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_adv_instruction_1_pipeline_en_5.4.2_3.0_1724027452360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_adv_instruction_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_adv_instruction_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_adv_instruction_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_adv_instruction_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_en.md new file mode 100644 index 00000000000000..4ca970f47c4deb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_bible_adv_instruction_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_bible_adv_instruction_2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_bible_adv_instruction_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bible_adv_instruction_2_en_5.4.2_3.0_1724042320713.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bible_adv_instruction_2_en_5.4.2_3.0_1724042320713.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_bible_adv_instruction_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_bible_adv_instruction_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_bible_adv_instruction_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_bible_adv_instruction_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_pipeline_en.md new file mode 100644 index 00000000000000..13cd9c68637317 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_bible_adv_instruction_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_bible_adv_instruction_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_bible_adv_instruction_2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_bible_adv_instruction_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bible_adv_instruction_2_pipeline_en_5.4.2_3.0_1724042453543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_bible_adv_instruction_2_pipeline_en_5.4.2_3.0_1724042453543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_bible_adv_instruction_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_bible_adv_instruction_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_bible_adv_instruction_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_bible_adv_instruction_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_en.md new file mode 100644 index 00000000000000..8fa7db94c0cee3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_flip_instruction_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_flip_instruction_0 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_flip_instruction_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_0_en_5.4.2_3.0_1724095567260.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_0_en_5.4.2_3.0_1724095567260.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_flip_instruction_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_flip_instruction_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_flip_instruction_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_flip_instruction_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_pipeline_en.md new file mode 100644 index 00000000000000..7d6adf7fe0d0bd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5large_tweet_emotion_flip_instruction_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_flip_instruction_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_flip_instruction_0_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_flip_instruction_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_0_pipeline_en_5.4.2_3.0_1724095714147.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_0_pipeline_en_5.4.2_3.0_1724095714147.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_flip_instruction_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_flip_instruction_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_flip_instruction_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_flip_instruction_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-t5v1_1_large_mnli_en.md b/docs/_posts/ahmedlone127/2024-08-19-t5v1_1_large_mnli_en.md new file mode 100644 index 00000000000000..d41021307f1105 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-t5v1_1_large_mnli_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5v1_1_large_mnli T5Transformer from pietrolesci +author: John Snow Labs +name: t5v1_1_large_mnli +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5v1_1_large_mnli` is a English model originally trained by pietrolesci. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5v1_1_large_mnli_en_5.4.2_3.0_1724054395651.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5v1_1_large_mnli_en_5.4.2_3.0_1724054395651.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5v1_1_large_mnli","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5v1_1_large_mnli", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5v1_1_large_mnli| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/pietrolesci/t5v1_1-large-mnli \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_en.md b/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_en.md new file mode 100644 index 00000000000000..6d05e2138a4495 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English task_a_flan_t5_large_rundi_3 T5Transformer from wanglab +author: John Snow Labs +name: task_a_flan_t5_large_rundi_3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`task_a_flan_t5_large_rundi_3` is a English model originally trained by wanglab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/task_a_flan_t5_large_rundi_3_en_5.4.2_3.0_1724071498533.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/task_a_flan_t5_large_rundi_3_en_5.4.2_3.0_1724071498533.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("task_a_flan_t5_large_rundi_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("task_a_flan_t5_large_rundi_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|task_a_flan_t5_large_rundi_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/wanglab/task-a-flan-t5-large-run-3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_pipeline_en.md new file mode 100644 index 00000000000000..69780e1d947380 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-task_a_flan_t5_large_rundi_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English task_a_flan_t5_large_rundi_3_pipeline pipeline T5Transformer from wanglab +author: John Snow Labs +name: task_a_flan_t5_large_rundi_3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`task_a_flan_t5_large_rundi_3_pipeline` is a English model originally trained by wanglab. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/task_a_flan_t5_large_rundi_3_pipeline_en_5.4.2_3.0_1724071636591.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/task_a_flan_t5_large_rundi_3_pipeline_en_5.4.2_3.0_1724071636591.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("task_a_flan_t5_large_rundi_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("task_a_flan_t5_large_rundi_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|task_a_flan_t5_large_rundi_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/wanglab/task-a-flan-t5-large-run-3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_en.md b/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_en.md new file mode 100644 index 00000000000000..2d48a2250e9b87 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English teabreac_t5_large_tatqa T5Transformer from StonyBrookNLP +author: John Snow Labs +name: teabreac_t5_large_tatqa +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`teabreac_t5_large_tatqa` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_tatqa_en_5.4.2_3.0_1724042106963.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_tatqa_en_5.4.2_3.0_1724042106963.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("teabreac_t5_large_tatqa","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("teabreac_t5_large_tatqa", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|teabreac_t5_large_tatqa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/StonyBrookNLP/teabreac-t5-large-tatqa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_pipeline_en.md new file mode 100644 index 00000000000000..8aa87346e826b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-teabreac_t5_large_tatqa_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English teabreac_t5_large_tatqa_pipeline pipeline T5Transformer from StonyBrookNLP +author: John Snow Labs +name: teabreac_t5_large_tatqa_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`teabreac_t5_large_tatqa_pipeline` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_tatqa_pipeline_en_5.4.2_3.0_1724042244761.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_tatqa_pipeline_en_5.4.2_3.0_1724042244761.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("teabreac_t5_large_tatqa_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("teabreac_t5_large_tatqa_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|teabreac_t5_large_tatqa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/StonyBrookNLP/teabreac-t5-large-tatqa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_en.md new file mode 100644 index 00000000000000..034ab17ebc99bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_finetuned_cnn_dailymail T5Transformer from CyrexPro +author: John Snow Labs +name: test_finetuned_cnn_dailymail +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_finetuned_cnn_dailymail` is a English model originally trained by CyrexPro. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_finetuned_cnn_dailymail_en_5.4.2_3.0_1724043053646.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_finetuned_cnn_dailymail_en_5.4.2_3.0_1724043053646.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_finetuned_cnn_dailymail","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_finetuned_cnn_dailymail", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_finetuned_cnn_dailymail| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/CyrexPro/Test-finetuned-cnn_dailymail \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_pipeline_en.md new file mode 100644 index 00000000000000..1862457d4c693d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_finetuned_cnn_dailymail_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_finetuned_cnn_dailymail_pipeline pipeline T5Transformer from CyrexPro +author: John Snow Labs +name: test_finetuned_cnn_dailymail_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_finetuned_cnn_dailymail_pipeline` is a English model originally trained by CyrexPro. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_finetuned_cnn_dailymail_pipeline_en_5.4.2_3.0_1724043070248.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_finetuned_cnn_dailymail_pipeline_en_5.4.2_3.0_1724043070248.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_finetuned_cnn_dailymail_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_finetuned_cnn_dailymail_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_finetuned_cnn_dailymail_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.5 MB| + +## References + +https://huggingface.co/CyrexPro/Test-finetuned-cnn_dailymail + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model3_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model3_en.md new file mode 100644 index 00000000000000..ff7c6a8c502cae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_model3 T5Transformer from atulxop +author: John Snow Labs +name: test_model3 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model3` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model3_en_5.4.2_3.0_1724087691645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model3_en_5.4.2_3.0_1724087691645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_model3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_model3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/atulxop/test_model3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model3_pipeline_en.md new file mode 100644 index 00000000000000..cc9bc913dd60a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_model3_pipeline pipeline T5Transformer from atulxop +author: John Snow Labs +name: test_model3_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model3_pipeline` is a English model originally trained by atulxop. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model3_pipeline_en_5.4.2_3.0_1724087708586.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model3_pipeline_en_5.4.2_3.0_1724087708586.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_model3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_model3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.9 MB| + +## References + +https://huggingface.co/atulxop/test_model3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_en.md new file mode 100644 index 00000000000000..a7190b543a8958 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_model_1e_4_30e T5Transformer from Yugratna +author: John Snow Labs +name: test_model_1e_4_30e +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_1e_4_30e` is a English model originally trained by Yugratna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_1e_4_30e_en_5.4.2_3.0_1724049522019.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_1e_4_30e_en_5.4.2_3.0_1724049522019.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_model_1e_4_30e","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_model_1e_4_30e", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_1e_4_30e| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.7 MB| + +## References + +https://huggingface.co/Yugratna/test_model_1e_4_30E \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_pipeline_en.md new file mode 100644 index 00000000000000..85e63161c56421 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_30e_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_model_1e_4_30e_pipeline pipeline T5Transformer from Yugratna +author: John Snow Labs +name: test_model_1e_4_30e_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_1e_4_30e_pipeline` is a English model originally trained by Yugratna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_1e_4_30e_pipeline_en_5.4.2_3.0_1724049541342.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_1e_4_30e_pipeline_en_5.4.2_3.0_1724049541342.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_model_1e_4_30e_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_model_1e_4_30e_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_1e_4_30e_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.7 MB| + +## References + +https://huggingface.co/Yugratna/test_model_1e_4_30E + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_en.md new file mode 100644 index 00000000000000..d97cf659021831 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_model_1e_4_50e T5Transformer from Yugratna +author: John Snow Labs +name: test_model_1e_4_50e +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_1e_4_50e` is a English model originally trained by Yugratna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_1e_4_50e_en_5.4.2_3.0_1724063803052.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_1e_4_50e_en_5.4.2_3.0_1724063803052.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_model_1e_4_50e","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_model_1e_4_50e", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_1e_4_50e| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.1 MB| + +## References + +https://huggingface.co/Yugratna/test_model_1e_4_50E \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_pipeline_en.md new file mode 100644 index 00000000000000..b71db6b74d0fd1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_1e_4_50e_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_model_1e_4_50e_pipeline pipeline T5Transformer from Yugratna +author: John Snow Labs +name: test_model_1e_4_50e_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_1e_4_50e_pipeline` is a English model originally trained by Yugratna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_1e_4_50e_pipeline_en_5.4.2_3.0_1724063819647.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_1e_4_50e_pipeline_en_5.4.2_3.0_1724063819647.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_model_1e_4_50e_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_model_1e_4_50e_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_1e_4_50e_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.1 MB| + +## References + +https://huggingface.co/Yugratna/test_model_1e_4_50E + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_en.md new file mode 100644 index 00000000000000..051b5de8c99367 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_model_ptt5_1_savinghf T5Transformer from godoyj +author: John Snow Labs +name: test_model_ptt5_1_savinghf +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_ptt5_1_savinghf` is a English model originally trained by godoyj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_ptt5_1_savinghf_en_5.4.2_3.0_1724072127565.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_ptt5_1_savinghf_en_5.4.2_3.0_1724072127565.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_model_ptt5_1_savinghf","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_model_ptt5_1_savinghf", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_ptt5_1_savinghf| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|978.8 MB| + +## References + +https://huggingface.co/godoyj/test-model-ptt5-1-savinghf \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_pipeline_en.md new file mode 100644 index 00000000000000..38c83a0a2ae684 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_model_ptt5_1_savinghf_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_model_ptt5_1_savinghf_pipeline pipeline T5Transformer from godoyj +author: John Snow Labs +name: test_model_ptt5_1_savinghf_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_model_ptt5_1_savinghf_pipeline` is a English model originally trained by godoyj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_model_ptt5_1_savinghf_pipeline_en_5.4.2_3.0_1724072185934.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_model_ptt5_1_savinghf_pipeline_en_5.4.2_3.0_1724072185934.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_model_ptt5_1_savinghf_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_model_ptt5_1_savinghf_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_model_ptt5_1_savinghf_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|978.8 MB| + +## References + +https://huggingface.co/godoyj/test-model-ptt5-1-savinghf + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_en.md new file mode 100644 index 00000000000000..0c8c1efc4f86d2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_sum_abs_t5_small_wasa_norwegian_stops T5Transformer from InfinityC +author: John Snow Labs +name: test_sum_abs_t5_small_wasa_norwegian_stops +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_sum_abs_t5_small_wasa_norwegian_stops` is a English model originally trained by InfinityC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_norwegian_stops_en_5.4.2_3.0_1724109612519.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_norwegian_stops_en_5.4.2_3.0_1724109612519.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_sum_abs_t5_small_wasa_norwegian_stops","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_sum_abs_t5_small_wasa_norwegian_stops", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_sum_abs_t5_small_wasa_norwegian_stops| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|325.2 MB| + +## References + +https://huggingface.co/InfinityC/test_sum_abs_t5_small_wasa_no_stops \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_pipeline_en.md new file mode 100644 index 00000000000000..6dd1beb7781fd3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-test_sum_abs_t5_small_wasa_norwegian_stops_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_sum_abs_t5_small_wasa_norwegian_stops_pipeline pipeline T5Transformer from InfinityC +author: John Snow Labs +name: test_sum_abs_t5_small_wasa_norwegian_stops_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_sum_abs_t5_small_wasa_norwegian_stops_pipeline` is a English model originally trained by InfinityC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_norwegian_stops_pipeline_en_5.4.2_3.0_1724109634606.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_sum_abs_t5_small_wasa_norwegian_stops_pipeline_en_5.4.2_3.0_1724109634606.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_sum_abs_t5_small_wasa_norwegian_stops_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_sum_abs_t5_small_wasa_norwegian_stops_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_sum_abs_t5_small_wasa_norwegian_stops_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|325.2 MB| + +## References + +https://huggingface.co/InfinityC/test_sum_abs_t5_small_wasa_no_stops + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-testing_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-testing_model_en.md new file mode 100644 index 00000000000000..9dc0b1abffe128 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-testing_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English testing_model T5Transformer from eatlas +author: John Snow Labs +name: testing_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`testing_model` is a English model originally trained by eatlas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/testing_model_en_5.4.2_3.0_1724098920945.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/testing_model_en_5.4.2_3.0_1724098920945.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("testing_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("testing_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|testing_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|280.2 MB| + +## References + +https://huggingface.co/eatlas/testing_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-testing_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-testing_model_pipeline_en.md new file mode 100644 index 00000000000000..65d3bf51387e57 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-testing_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English testing_model_pipeline pipeline T5Transformer from eatlas +author: John Snow Labs +name: testing_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`testing_model_pipeline` is a English model originally trained by eatlas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/testing_model_pipeline_en_5.4.2_3.0_1724098950804.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/testing_model_pipeline_en_5.4.2_3.0_1724098950804.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("testing_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("testing_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|testing_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|280.2 MB| + +## References + +https://huggingface.co/eatlas/testing_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_en.md new file mode 100644 index 00000000000000..08930b732537ba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_exps_t5_20 T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_20 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_20` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_en_5.4.2_3.0_1724109479889.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_en_5.4.2_3.0_1724109479889.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_exps_t5_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_exps_t5_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|967.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_pipeline_en.md new file mode 100644 index 00000000000000..0c13a0f6cb9676 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_exps_t5_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_exps_t5_20_pipeline pipeline T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_20_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_20_pipeline` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_pipeline_en_5.4.2_3.0_1724109531198.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_pipeline_en_5.4.2_3.0_1724109531198.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_exps_t5_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_exps_t5_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|967.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_en.md new file mode 100644 index 00000000000000..1158216023573e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_shortening_model_v12 T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v12 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v12` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v12_en_5.4.2_3.0_1724040433263.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v12_en_5.4.2_3.0_1724040433263.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_shortening_model_v12","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_shortening_model_v12", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v12| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|334.8 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v12 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_pipeline_en.md new file mode 100644 index 00000000000000..916ffa1928c476 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v12_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_shortening_model_v12_pipeline pipeline T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v12_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v12_pipeline` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v12_pipeline_en_5.4.2_3.0_1724040450811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v12_pipeline_en_5.4.2_3.0_1724040450811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_shortening_model_v12_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_shortening_model_v12_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v12_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|334.8 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v12 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_en.md new file mode 100644 index 00000000000000..259ad80cb95240 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_shortening_model_v5 T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v5` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v5_en_5.4.2_3.0_1724063317273.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v5_en_5.4.2_3.0_1724063317273.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_shortening_model_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_shortening_model_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_pipeline_en.md new file mode 100644 index 00000000000000..7938e1e0fa697c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_shortening_model_v5_pipeline pipeline T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v5_pipeline` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v5_pipeline_en_5.4.2_3.0_1724063335944.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v5_pipeline_en_5.4.2_3.0_1724063335944.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_shortening_model_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_shortening_model_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_en.md new file mode 100644 index 00000000000000..d9d5b09cfee4b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_shortening_model_v78 T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v78 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v78` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v78_en_5.4.2_3.0_1724089769522.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v78_en_5.4.2_3.0_1724089769522.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_shortening_model_v78","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_shortening_model_v78", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v78| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.6 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v78 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_pipeline_en.md new file mode 100644 index 00000000000000..b260eead76db9c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-text_shortening_model_v78_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_shortening_model_v78_pipeline pipeline T5Transformer from ldos +author: John Snow Labs +name: text_shortening_model_v78_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_shortening_model_v78_pipeline` is a English model originally trained by ldos. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_shortening_model_v78_pipeline_en_5.4.2_3.0_1724089786640.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_shortening_model_v78_pipeline_en_5.4.2_3.0_1724089786640.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_shortening_model_v78_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_shortening_model_v78_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_shortening_model_v78_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.6 MB| + +## References + +https://huggingface.co/ldos/text_shortening_model_v78 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_en.md new file mode 100644 index 00000000000000..4055d900f49158 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English titlewave_t5_small T5Transformer from tennessejoyce +author: John Snow Labs +name: titlewave_t5_small +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`titlewave_t5_small` is a English model originally trained by tennessejoyce. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/titlewave_t5_small_en_5.4.2_3.0_1724048632763.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/titlewave_t5_small_en_5.4.2_3.0_1724048632763.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("titlewave_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("titlewave_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|titlewave_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.5 MB| + +## References + +https://huggingface.co/tennessejoyce/titlewave-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..e281b242a84e9e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-titlewave_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English titlewave_t5_small_pipeline pipeline T5Transformer from tennessejoyce +author: John Snow Labs +name: titlewave_t5_small_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`titlewave_t5_small_pipeline` is a English model originally trained by tennessejoyce. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/titlewave_t5_small_pipeline_en_5.4.2_3.0_1724048651863.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/titlewave_t5_small_pipeline_en_5.4.2_3.0_1724048651863.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("titlewave_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("titlewave_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|titlewave_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.5 MB| + +## References + +https://huggingface.co/tennessejoyce/titlewave-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_en.md b/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_en.md new file mode 100644 index 00000000000000..eb3f3e923cd98b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tmp_noise_t5_001 T5Transformer from jinfwhuang +author: John Snow Labs +name: tmp_noise_t5_001 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tmp_noise_t5_001` is a English model originally trained by jinfwhuang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tmp_noise_t5_001_en_5.4.2_3.0_1724058090655.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tmp_noise_t5_001_en_5.4.2_3.0_1724058090655.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tmp_noise_t5_001","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tmp_noise_t5_001", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tmp_noise_t5_001| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|274.2 MB| + +## References + +https://huggingface.co/jinfwhuang/tmp-noise-t5-001 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_pipeline_en.md new file mode 100644 index 00000000000000..02df28c4310600 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-tmp_noise_t5_001_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tmp_noise_t5_001_pipeline pipeline T5Transformer from jinfwhuang +author: John Snow Labs +name: tmp_noise_t5_001_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tmp_noise_t5_001_pipeline` is a English model originally trained by jinfwhuang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tmp_noise_t5_001_pipeline_en_5.4.2_3.0_1724058113980.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tmp_noise_t5_001_pipeline_en_5.4.2_3.0_1724058113980.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tmp_noise_t5_001_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tmp_noise_t5_001_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tmp_noise_t5_001_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|274.2 MB| + +## References + +https://huggingface.co/jinfwhuang/tmp-noise-t5-001 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-touring2_en.md b/docs/_posts/ahmedlone127/2024-08-19-touring2_en.md new file mode 100644 index 00000000000000..793a39cffeb099 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-touring2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English touring2 T5Transformer from acrowth +author: John Snow Labs +name: touring2 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`touring2` is a English model originally trained by acrowth. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/touring2_en_5.4.2_3.0_1724102878926.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/touring2_en_5.4.2_3.0_1724102878926.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("touring2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("touring2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|touring2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acrowth/touring2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-touring2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-touring2_pipeline_en.md new file mode 100644 index 00000000000000..6eb413c4d901d5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-touring2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English touring2_pipeline pipeline T5Transformer from acrowth +author: John Snow Labs +name: touring2_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`touring2_pipeline` is a English model originally trained by acrowth. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/touring2_pipeline_en_5.4.2_3.0_1724102926487.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/touring2_pipeline_en_5.4.2_3.0_1724102926487.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("touring2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("touring2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|touring2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/acrowth/touring2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-train_t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-train_t5_en.md new file mode 100644 index 00000000000000..764bd7df9bc08c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-train_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English train_t5 T5Transformer from santyzenith +author: John Snow Labs +name: train_t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`train_t5` is a English model originally trained by santyzenith. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/train_t5_en_5.4.2_3.0_1724054460658.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/train_t5_en_5.4.2_3.0_1724054460658.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("train_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("train_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|train_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.4 MB| + +## References + +https://huggingface.co/santyzenith/train_t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-train_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-train_t5_pipeline_en.md new file mode 100644 index 00000000000000..0050fe97a2ac1b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-train_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English train_t5_pipeline pipeline T5Transformer from santyzenith +author: John Snow Labs +name: train_t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`train_t5_pipeline` is a English model originally trained by santyzenith. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/train_t5_pipeline_en_5.4.2_3.0_1724054476878.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/train_t5_pipeline_en_5.4.2_3.0_1724054476878.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("train_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("train_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|train_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.4 MB| + +## References + +https://huggingface.co/santyzenith/train_t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-transforemr_en.md b/docs/_posts/ahmedlone127/2024-08-19-transforemr_en.md new file mode 100644 index 00000000000000..ac189b9c2c38bb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-transforemr_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English transforemr T5Transformer from mins0o0 +author: John Snow Labs +name: transforemr +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`transforemr` is a English model originally trained by mins0o0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/transforemr_en_5.4.2_3.0_1724043766789.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/transforemr_en_5.4.2_3.0_1724043766789.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("transforemr","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("transforemr", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|transforemr| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.5 MB| + +## References + +https://huggingface.co/mins0o0/transforemr \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-transforemr_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-transforemr_pipeline_en.md new file mode 100644 index 00000000000000..b619a3ce34cbd6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-transforemr_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English transforemr_pipeline pipeline T5Transformer from mins0o0 +author: John Snow Labs +name: transforemr_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`transforemr_pipeline` is a English model originally trained by mins0o0. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/transforemr_pipeline_en_5.4.2_3.0_1724043784659.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/transforemr_pipeline_en_5.4.2_3.0_1724043784659.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("transforemr_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("transforemr_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|transforemr_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.5 MB| + +## References + +https://huggingface.co/mins0o0/transforemr + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_en.md new file mode 100644 index 00000000000000..7b48e8ee439ffd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English tuple_1k_t5 T5Transformer from tombenj +author: John Snow Labs +name: tuple_1k_t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tuple_1k_t5` is a English model originally trained by tombenj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tuple_1k_t5_en_5.4.2_3.0_1724100197298.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tuple_1k_t5_en_5.4.2_3.0_1724100197298.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("tuple_1k_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("tuple_1k_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tuple_1k_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|891.8 MB| + +## References + +https://huggingface.co/tombenj/tuple-1k-t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_pipeline_en.md new file mode 100644 index 00000000000000..5d3a9a951dad6f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-tuple_1k_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English tuple_1k_t5_pipeline pipeline T5Transformer from tombenj +author: John Snow Labs +name: tuple_1k_t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`tuple_1k_t5_pipeline` is a English model originally trained by tombenj. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/tuple_1k_t5_pipeline_en_5.4.2_3.0_1724100279360.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/tuple_1k_t5_pipeline_en_5.4.2_3.0_1724100279360.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("tuple_1k_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("tuple_1k_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|tuple_1k_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|891.8 MB| + +## References + +https://huggingface.co/tombenj/tuple-1k-t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_en.md b/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_en.md new file mode 100644 index 00000000000000..8b27e1dd582e0a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English two_variable_large_model T5Transformer from sammanamgain +author: John Snow Labs +name: two_variable_large_model +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`two_variable_large_model` is a English model originally trained by sammanamgain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/two_variable_large_model_en_5.4.2_3.0_1724096874418.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/two_variable_large_model_en_5.4.2_3.0_1724096874418.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("two_variable_large_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("two_variable_large_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|two_variable_large_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/sammanamgain/Two_variable_large_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_pipeline_en.md new file mode 100644 index 00000000000000..b9e354d9824f71 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-two_variable_large_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English two_variable_large_model_pipeline pipeline T5Transformer from sammanamgain +author: John Snow Labs +name: two_variable_large_model_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`two_variable_large_model_pipeline` is a English model originally trained by sammanamgain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/two_variable_large_model_pipeline_en_5.4.2_3.0_1724097021651.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/two_variable_large_model_pipeline_en_5.4.2_3.0_1724097021651.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("two_variable_large_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("two_variable_large_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|two_variable_large_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/sammanamgain/Two_variable_large_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_en.md b/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_en.md new file mode 100644 index 00000000000000..472e543b00051b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English uitviquad_t5 T5Transformer from TGiang +author: John Snow Labs +name: uitviquad_t5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`uitviquad_t5` is a English model originally trained by TGiang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/uitviquad_t5_en_5.4.2_3.0_1724108399110.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/uitviquad_t5_en_5.4.2_3.0_1724108399110.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("uitviquad_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("uitviquad_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|uitviquad_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|982.3 MB| + +## References + +https://huggingface.co/TGiang/uitviquad_t5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_pipeline_en.md new file mode 100644 index 00000000000000..635ea47346a5f3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-uitviquad_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English uitviquad_t5_pipeline pipeline T5Transformer from TGiang +author: John Snow Labs +name: uitviquad_t5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`uitviquad_t5_pipeline` is a English model originally trained by TGiang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/uitviquad_t5_pipeline_en_5.4.2_3.0_1724108453516.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/uitviquad_t5_pipeline_en_5.4.2_3.0_1724108453516.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("uitviquad_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("uitviquad_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|uitviquad_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|982.3 MB| + +## References + +https://huggingface.co/TGiang/uitviquad_t5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_nl.md b/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_nl.md new file mode 100644 index 00000000000000..10155d0cc72ba6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_nl.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Dutch, Flemish ul2_base_nl36_dutch T5Transformer from yhavinga +author: John Snow Labs +name: ul2_base_nl36_dutch +date: 2024-08-19 +tags: [nl, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: nl +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ul2_base_nl36_dutch` is a Dutch, Flemish model originally trained by yhavinga. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ul2_base_nl36_dutch_nl_5.4.2_3.0_1724080921554.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ul2_base_nl36_dutch_nl_5.4.2_3.0_1724080921554.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ul2_base_nl36_dutch","nl") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ul2_base_nl36_dutch", "nl") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ul2_base_nl36_dutch| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|nl| +|Size:|3.1 GB| + +## References + +https://huggingface.co/yhavinga/ul2-base-nl36-dutch \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_pipeline_nl.md b/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_pipeline_nl.md new file mode 100644 index 00000000000000..8c0a489e6a9cc3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-ul2_base_nl36_dutch_pipeline_nl.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Dutch, Flemish ul2_base_nl36_dutch_pipeline pipeline T5Transformer from yhavinga +author: John Snow Labs +name: ul2_base_nl36_dutch_pipeline +date: 2024-08-19 +tags: [nl, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: nl +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ul2_base_nl36_dutch_pipeline` is a Dutch, Flemish model originally trained by yhavinga. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ul2_base_nl36_dutch_pipeline_nl_5.4.2_3.0_1724081061865.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ul2_base_nl36_dutch_pipeline_nl_5.4.2_3.0_1724081061865.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ul2_base_nl36_dutch_pipeline", lang = "nl") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ul2_base_nl36_dutch_pipeline", lang = "nl") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ul2_base_nl36_dutch_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|nl| +|Size:|3.1 GB| + +## References + +https://huggingface.co/yhavinga/ul2-base-nl36-dutch + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-venkatt5_en.md b/docs/_posts/ahmedlone127/2024-08-19-venkatt5_en.md new file mode 100644 index 00000000000000..7868034663387f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-venkatt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English venkatt5 T5Transformer from venky26 +author: John Snow Labs +name: venkatt5 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`venkatt5` is a English model originally trained by venky26. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/venkatt5_en_5.4.2_3.0_1724042793082.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/venkatt5_en_5.4.2_3.0_1724042793082.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("venkatt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("venkatt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|venkatt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|988.7 MB| + +## References + +https://huggingface.co/venky26/VenkatT5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-venkatt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-venkatt5_pipeline_en.md new file mode 100644 index 00000000000000..9d819939834644 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-venkatt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English venkatt5_pipeline pipeline T5Transformer from venky26 +author: John Snow Labs +name: venkatt5_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`venkatt5_pipeline` is a English model originally trained by venky26. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/venkatt5_pipeline_en_5.4.2_3.0_1724042848543.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/venkatt5_pipeline_en_5.4.2_3.0_1724042848543.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("venkatt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("venkatt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|venkatt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|988.7 MB| + +## References + +https://huggingface.co/venky26/VenkatT5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_en.md b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_en.md new file mode 100644 index 00000000000000..a29144613f19a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietnamese_english_envit5_translation_doc_train T5Transformer from hungphongtrn +author: John Snow Labs +name: vietnamese_english_envit5_translation_doc_train +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_english_envit5_translation_doc_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_doc_train_en_5.4.2_3.0_1724042343732.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_doc_train_en_5.4.2_3.0_1724042343732.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietnamese_english_envit5_translation_doc_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietnamese_english_envit5_translation_doc_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_english_envit5_translation_doc_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vi_en_envit5-translation_doc_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_pipeline_en.md new file mode 100644 index 00000000000000..e4f4b0f4eac71f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_english_envit5_translation_doc_train_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietnamese_english_envit5_translation_doc_train_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: vietnamese_english_envit5_translation_doc_train_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_english_envit5_translation_doc_train_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_doc_train_pipeline_en_5.4.2_3.0_1724042415882.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_english_envit5_translation_doc_train_pipeline_en_5.4.2_3.0_1724042415882.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietnamese_english_envit5_translation_doc_train_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietnamese_english_envit5_translation_doc_train_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_english_envit5_translation_doc_train_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vi_en_envit5-translation_doc_train + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-vietnamese_t5_base_finetune_rewriter_6_epochs_en.md b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_t5_base_finetune_rewriter_6_epochs_en.md new file mode 100644 index 00000000000000..dac02b35c3b67d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-vietnamese_t5_base_finetune_rewriter_6_epochs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter_6_epochs T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter_6_epochs +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter_6_epochs` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_6_epochs_en_5.4.2_3.0_1724111981918.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_6_epochs_en_5.4.2_3.0_1724111981918.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter_6_epochs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietnamese_t5_base_finetune_rewriter_6_epochs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter_6_epochs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter-6-epochs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_en.md b/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_en.md new file mode 100644 index 00000000000000..f38baa25864048 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English xor_tydi_doctquery_mt5_base T5Transformer from ielabgroup +author: John Snow Labs +name: xor_tydi_doctquery_mt5_base +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xor_tydi_doctquery_mt5_base` is a English model originally trained by ielabgroup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xor_tydi_doctquery_mt5_base_en_5.4.2_3.0_1724028252200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xor_tydi_doctquery_mt5_base_en_5.4.2_3.0_1724028252200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("xor_tydi_doctquery_mt5_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("xor_tydi_doctquery_mt5_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xor_tydi_doctquery_mt5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/ielabgroup/xor-tydi-docTquery-mt5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_pipeline_en.md new file mode 100644 index 00000000000000..88e22c3a832f24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-xor_tydi_doctquery_mt5_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English xor_tydi_doctquery_mt5_base_pipeline pipeline T5Transformer from ielabgroup +author: John Snow Labs +name: xor_tydi_doctquery_mt5_base_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xor_tydi_doctquery_mt5_base_pipeline` is a English model originally trained by ielabgroup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xor_tydi_doctquery_mt5_base_pipeline_en_5.4.2_3.0_1724028508282.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xor_tydi_doctquery_mt5_base_pipeline_en_5.4.2_3.0_1724028508282.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("xor_tydi_doctquery_mt5_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("xor_tydi_doctquery_mt5_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xor_tydi_doctquery_mt5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/ielabgroup/xor-tydi-docTquery-mt5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md b/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md new file mode 100644 index 00000000000000..f7db299bdf288d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6 T5Transformer from KingKazma +author: John Snow Labs +name: xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en_5.4.2_3.0_1724034505880.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en_5.4.2_3.0_1724034505880.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|301.4 MB| + +## References + +https://huggingface.co/KingKazma/xsum_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md new file mode 100644 index 00000000000000..71a2c781aac64c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline pipeline T5Transformer from KingKazma +author: John Snow Labs +name: xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en_5.4.2_3.0_1724034533625.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en_5.4.2_3.0_1724034533625.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|xsum_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|301.4 MB| + +## References + +https://huggingface.co/KingKazma/xsum_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_en.md new file mode 100644 index 00000000000000..c68bcbc101baf4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English yelp_polarity_t5_base_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: yelp_polarity_t5_base_seed_1 +date: 2024-08-19 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`yelp_polarity_t5_base_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_1_en_5.4.2_3.0_1724083389553.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_1_en_5.4.2_3.0_1724083389553.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("yelp_polarity_t5_base_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("yelp_polarity_t5_base_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|yelp_polarity_t5_base_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/utahnlp/yelp_polarity_t5-base_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..fcb82b537dbcea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-19-yelp_polarity_t5_base_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English yelp_polarity_t5_base_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: yelp_polarity_t5_base_seed_1_pipeline +date: 2024-08-19 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`yelp_polarity_t5_base_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724083440793.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/yelp_polarity_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724083440793.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("yelp_polarity_t5_base_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("yelp_polarity_t5_base_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|yelp_polarity_t5_base_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/utahnlp/yelp_polarity_t5-base_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-20240213_3_en.md b/docs/_posts/ahmedlone127/2024-08-20-20240213_3_en.md new file mode 100644 index 00000000000000..dc6d1178de878d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-20240213_3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English 20240213_3 T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_3` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_3_en_5.4.2_3.0_1724129813106.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_3_en_5.4.2_3.0_1724129813106.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("20240213_3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("20240213_3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/picas9dan/20240213_3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-20240213_3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-20240213_3_pipeline_en.md new file mode 100644 index 00000000000000..d69075111c22a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-20240213_3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English 20240213_3_pipeline pipeline T5Transformer from picas9dan +author: John Snow Labs +name: 20240213_3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`20240213_3_pipeline` is a English model originally trained by picas9dan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/20240213_3_pipeline_en_5.4.2_3.0_1724129860769.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/20240213_3_pipeline_en_5.4.2_3.0_1724129860769.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("20240213_3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("20240213_3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|20240213_3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/picas9dan/20240213_3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_en.md b/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_en.md new file mode 100644 index 00000000000000..ce4f42343b0871 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English afrimt5_english_swahili_macrolanguage_news T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_english_swahili_macrolanguage_news +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_english_swahili_macrolanguage_news` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_english_swahili_macrolanguage_news_en_5.4.2_3.0_1724128801669.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_english_swahili_macrolanguage_news_en_5.4.2_3.0_1724128801669.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("afrimt5_english_swahili_macrolanguage_news","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("afrimt5_english_swahili_macrolanguage_news", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_english_swahili_macrolanguage_news| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_en_swa_news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_pipeline_en.md new file mode 100644 index 00000000000000..a84f2aada2840f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-afrimt5_english_swahili_macrolanguage_news_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English afrimt5_english_swahili_macrolanguage_news_pipeline pipeline T5Transformer from masakhane +author: John Snow Labs +name: afrimt5_english_swahili_macrolanguage_news_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`afrimt5_english_swahili_macrolanguage_news_pipeline` is a English model originally trained by masakhane. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/afrimt5_english_swahili_macrolanguage_news_pipeline_en_5.4.2_3.0_1724128945633.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/afrimt5_english_swahili_macrolanguage_news_pipeline_en_5.4.2_3.0_1724128945633.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("afrimt5_english_swahili_macrolanguage_news_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("afrimt5_english_swahili_macrolanguage_news_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|afrimt5_english_swahili_macrolanguage_news_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.8 GB| + +## References + +https://huggingface.co/masakhane/afrimt5_en_swa_news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_en.md b/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_en.md new file mode 100644 index 00000000000000..8321d290645f71 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English all_mt5_base_15_wikisql_norwegian_sch T5Transformer from e22vvb +author: John Snow Labs +name: all_mt5_base_15_wikisql_norwegian_sch +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_mt5_base_15_wikisql_norwegian_sch` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_mt5_base_15_wikisql_norwegian_sch_en_5.4.2_3.0_1724134371166.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_mt5_base_15_wikisql_norwegian_sch_en_5.4.2_3.0_1724134371166.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("all_mt5_base_15_wikisql_norwegian_sch","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("all_mt5_base_15_wikisql_norwegian_sch", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_mt5_base_15_wikisql_norwegian_sch| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/e22vvb/ALL_mt5-base_15_wikiSQL_no_sch \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_pipeline_en.md new file mode 100644 index 00000000000000..929b3800118866 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-all_mt5_base_15_wikisql_norwegian_sch_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English all_mt5_base_15_wikisql_norwegian_sch_pipeline pipeline T5Transformer from e22vvb +author: John Snow Labs +name: all_mt5_base_15_wikisql_norwegian_sch_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`all_mt5_base_15_wikisql_norwegian_sch_pipeline` is a English model originally trained by e22vvb. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/all_mt5_base_15_wikisql_norwegian_sch_pipeline_en_5.4.2_3.0_1724134642723.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/all_mt5_base_15_wikisql_norwegian_sch_pipeline_en_5.4.2_3.0_1724134642723.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("all_mt5_base_15_wikisql_norwegian_sch_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("all_mt5_base_15_wikisql_norwegian_sch_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|all_mt5_base_15_wikisql_norwegian_sch_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/e22vvb/ALL_mt5-base_15_wikiSQL_no_sch + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_en.md b/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_en.md new file mode 100644 index 00000000000000..a36e2086bb7e25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English alqalam_finetuned_mmj_withxlsumvalid T5Transformer from Grinta-king +author: John Snow Labs +name: alqalam_finetuned_mmj_withxlsumvalid +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`alqalam_finetuned_mmj_withxlsumvalid` is a English model originally trained by Grinta-king. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/alqalam_finetuned_mmj_withxlsumvalid_en_5.4.2_3.0_1724124885022.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/alqalam_finetuned_mmj_withxlsumvalid_en_5.4.2_3.0_1724124885022.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("alqalam_finetuned_mmj_withxlsumvalid","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("alqalam_finetuned_mmj_withxlsumvalid", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|alqalam_finetuned_mmj_withxlsumvalid| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/Grinta-king/AlQalam-finetuned-mmj-withXlsumValid \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_pipeline_en.md new file mode 100644 index 00000000000000..e83faaa46d56b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-alqalam_finetuned_mmj_withxlsumvalid_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English alqalam_finetuned_mmj_withxlsumvalid_pipeline pipeline T5Transformer from Grinta-king +author: John Snow Labs +name: alqalam_finetuned_mmj_withxlsumvalid_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`alqalam_finetuned_mmj_withxlsumvalid_pipeline` is a English model originally trained by Grinta-king. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/alqalam_finetuned_mmj_withxlsumvalid_pipeline_en_5.4.2_3.0_1724124965928.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/alqalam_finetuned_mmj_withxlsumvalid_pipeline_en_5.4.2_3.0_1724124965928.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("alqalam_finetuned_mmj_withxlsumvalid_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("alqalam_finetuned_mmj_withxlsumvalid_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|alqalam_finetuned_mmj_withxlsumvalid_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/Grinta-king/AlQalam-finetuned-mmj-withXlsumValid + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_en.md b/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_en.md new file mode 100644 index 00000000000000..ef462cd5f03a50 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ans_scorer_flan_large T5Transformer from divers +author: John Snow Labs +name: ans_scorer_flan_large +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ans_scorer_flan_large` is a English model originally trained by divers. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ans_scorer_flan_large_en_5.4.2_3.0_1724114065010.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ans_scorer_flan_large_en_5.4.2_3.0_1724114065010.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ans_scorer_flan_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ans_scorer_flan_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ans_scorer_flan_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/divers/ans-scorer-flan-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_pipeline_en.md new file mode 100644 index 00000000000000..28be37ca39893d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ans_scorer_flan_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ans_scorer_flan_large_pipeline pipeline T5Transformer from divers +author: John Snow Labs +name: ans_scorer_flan_large_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ans_scorer_flan_large_pipeline` is a English model originally trained by divers. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ans_scorer_flan_large_pipeline_en_5.4.2_3.0_1724114203726.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ans_scorer_flan_large_pipeline_en_5.4.2_3.0_1724114203726.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ans_scorer_flan_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ans_scorer_flan_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ans_scorer_flan_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/divers/ans-scorer-flan-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_en.md b/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_en.md new file mode 100644 index 00000000000000..a4293f4ce670fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English arat5_msa_small_finetuned_xlsum_arabic T5Transformer from Osame1 +author: John Snow Labs +name: arat5_msa_small_finetuned_xlsum_arabic +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arat5_msa_small_finetuned_xlsum_arabic` is a English model originally trained by Osame1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arat5_msa_small_finetuned_xlsum_arabic_en_5.4.2_3.0_1724120430817.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arat5_msa_small_finetuned_xlsum_arabic_en_5.4.2_3.0_1724120430817.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("arat5_msa_small_finetuned_xlsum_arabic","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("arat5_msa_small_finetuned_xlsum_arabic", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arat5_msa_small_finetuned_xlsum_arabic| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|828.7 MB| + +## References + +https://huggingface.co/Osame1/AraT5-msa-small-finetuned-xlsum-ar \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_pipeline_en.md new file mode 100644 index 00000000000000..dbc57f0f029372 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-arat5_msa_small_finetuned_xlsum_arabic_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English arat5_msa_small_finetuned_xlsum_arabic_pipeline pipeline T5Transformer from Osame1 +author: John Snow Labs +name: arat5_msa_small_finetuned_xlsum_arabic_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`arat5_msa_small_finetuned_xlsum_arabic_pipeline` is a English model originally trained by Osame1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/arat5_msa_small_finetuned_xlsum_arabic_pipeline_en_5.4.2_3.0_1724120475008.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/arat5_msa_small_finetuned_xlsum_arabic_pipeline_en_5.4.2_3.0_1724120475008.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("arat5_msa_small_finetuned_xlsum_arabic_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("arat5_msa_small_finetuned_xlsum_arabic_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|arat5_msa_small_finetuned_xlsum_arabic_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|828.7 MB| + +## References + +https://huggingface.co/Osame1/AraT5-msa-small-finetuned-xlsum-ar + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_en.md b/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_en.md new file mode 100644 index 00000000000000..13acfc96ad319c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English augmented_nodes_without_edge_label_sentence_level_t5_run3 T5Transformer from sheoran95 +author: John Snow Labs +name: augmented_nodes_without_edge_label_sentence_level_t5_run3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`augmented_nodes_without_edge_label_sentence_level_t5_run3` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/augmented_nodes_without_edge_label_sentence_level_t5_run3_en_5.4.2_3.0_1724117239160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/augmented_nodes_without_edge_label_sentence_level_t5_run3_en_5.4.2_3.0_1724117239160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("augmented_nodes_without_edge_label_sentence_level_t5_run3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("augmented_nodes_without_edge_label_sentence_level_t5_run3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|augmented_nodes_without_edge_label_sentence_level_t5_run3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|328.2 MB| + +## References + +https://huggingface.co/sheoran95/augmented_nodes_without_edge_label_sentence_level_T5_run3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline_en.md new file mode 100644 index 00000000000000..b84bf8f257d5c2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline_en_5.4.2_3.0_1724117257851.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline_en_5.4.2_3.0_1724117257851.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|augmented_nodes_without_edge_label_sentence_level_t5_run3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|328.2 MB| + +## References + +https://huggingface.co/sheoran95/augmented_nodes_without_edge_label_sentence_level_T5_run3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_en.md b/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_en.md new file mode 100644 index 00000000000000..79a2e05dbb4362 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bangla_para_v3_210000 T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_210000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_210000` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_210000_en_5.4.2_3.0_1724141363359.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_210000_en_5.4.2_3.0_1724141363359.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bangla_para_v3_210000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bangla_para_v3_210000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_210000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-210000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_pipeline_en.md new file mode 100644 index 00000000000000..dbff37d203fd37 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bangla_para_v3_210000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bangla_para_v3_210000_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: bangla_para_v3_210000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bangla_para_v3_210000_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bangla_para_v3_210000_pipeline_en_5.4.2_3.0_1724141409932.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bangla_para_v3_210000_pipeline_en_5.4.2_3.0_1724141409932.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bangla_para_v3_210000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bangla_para_v3_210000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bangla_para_v3_210000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mHossain/bangla-para-v3-210000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_en.md b/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_en.md new file mode 100644 index 00000000000000..1b802f439c5a22 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English banglat5_finetuned_headlinebt5_1000 T5Transformer from mdosama39 +author: John Snow Labs +name: banglat5_finetuned_headlinebt5_1000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5_finetuned_headlinebt5_1000` is a English model originally trained by mdosama39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_finetuned_headlinebt5_1000_en_5.4.2_3.0_1724131038965.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_finetuned_headlinebt5_1000_en_5.4.2_3.0_1724131038965.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("banglat5_finetuned_headlinebt5_1000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("banglat5_finetuned_headlinebt5_1000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5_finetuned_headlinebt5_1000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|960.9 MB| + +## References + +https://huggingface.co/mdosama39/banglat5-finetuned-headlineBT5_1000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_pipeline_en.md new file mode 100644 index 00000000000000..60b7d44edcc040 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-banglat5_finetuned_headlinebt5_1000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English banglat5_finetuned_headlinebt5_1000_pipeline pipeline T5Transformer from mdosama39 +author: John Snow Labs +name: banglat5_finetuned_headlinebt5_1000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`banglat5_finetuned_headlinebt5_1000_pipeline` is a English model originally trained by mdosama39. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/banglat5_finetuned_headlinebt5_1000_pipeline_en_5.4.2_3.0_1724131095963.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/banglat5_finetuned_headlinebt5_1000_pipeline_en_5.4.2_3.0_1724131095963.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("banglat5_finetuned_headlinebt5_1000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("banglat5_finetuned_headlinebt5_1000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|banglat5_finetuned_headlinebt5_1000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|960.9 MB| + +## References + +https://huggingface.co/mdosama39/banglat5-finetuned-headlineBT5_1000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_en.md b/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_en.md new file mode 100644 index 00000000000000..7f1d13b770f2dc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bengali_news_article_summarization_mt5 T5Transformer from fahad1770 +author: John Snow Labs +name: bengali_news_article_summarization_mt5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bengali_news_article_summarization_mt5` is a English model originally trained by fahad1770. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bengali_news_article_summarization_mt5_en_5.4.2_3.0_1724118615205.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bengali_news_article_summarization_mt5_en_5.4.2_3.0_1724118615205.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bengali_news_article_summarization_mt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bengali_news_article_summarization_mt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bengali_news_article_summarization_mt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/fahad1770/bengali_news_article_summarization_mt5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_pipeline_en.md new file mode 100644 index 00000000000000..9a88e680dfc719 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bengali_news_article_summarization_mt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bengali_news_article_summarization_mt5_pipeline pipeline T5Transformer from fahad1770 +author: John Snow Labs +name: bengali_news_article_summarization_mt5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bengali_news_article_summarization_mt5_pipeline` is a English model originally trained by fahad1770. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bengali_news_article_summarization_mt5_pipeline_en_5.4.2_3.0_1724118709914.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bengali_news_article_summarization_mt5_pipeline_en_5.4.2_3.0_1724118709914.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bengali_news_article_summarization_mt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bengali_news_article_summarization_mt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bengali_news_article_summarization_mt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/fahad1770/bengali_news_article_summarization_mt5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_en.md b/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_en.md new file mode 100644 index 00000000000000..7e9f9c73fb0e66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English bikes_t5_small_12 T5Transformer from neal61 +author: John Snow Labs +name: bikes_t5_small_12 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small_12` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_12_en_5.4.2_3.0_1724137851217.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_12_en_5.4.2_3.0_1724137851217.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("bikes_t5_small_12","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("bikes_t5_small_12", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small_12| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.2 MB| + +## References + +https://huggingface.co/neal61/bikes-t5-small-12 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_pipeline_en.md new file mode 100644 index 00000000000000..28c23f181ab908 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-bikes_t5_small_12_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English bikes_t5_small_12_pipeline pipeline T5Transformer from neal61 +author: John Snow Labs +name: bikes_t5_small_12_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`bikes_t5_small_12_pipeline` is a English model originally trained by neal61. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/bikes_t5_small_12_pipeline_en_5.4.2_3.0_1724137868112.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/bikes_t5_small_12_pipeline_en_5.4.2_3.0_1724137868112.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("bikes_t5_small_12_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("bikes_t5_small_12_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|bikes_t5_small_12_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.2 MB| + +## References + +https://huggingface.co/neal61/bikes-t5-small-12 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_en.md b/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_en.md new file mode 100644 index 00000000000000..aec79d2a4c1f7a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English boolq_t5_base_seed_1 T5Transformer from utahnlp +author: John Snow Labs +name: boolq_t5_base_seed_1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boolq_t5_base_seed_1` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_1_en_5.4.2_3.0_1724148872727.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_1_en_5.4.2_3.0_1724148872727.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("boolq_t5_base_seed_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("boolq_t5_base_seed_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boolq_t5_base_seed_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|963.9 MB| + +## References + +https://huggingface.co/utahnlp/boolq_t5-base_seed-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_pipeline_en.md new file mode 100644 index 00000000000000..a3b380b09812a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-boolq_t5_base_seed_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English boolq_t5_base_seed_1_pipeline pipeline T5Transformer from utahnlp +author: John Snow Labs +name: boolq_t5_base_seed_1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`boolq_t5_base_seed_1_pipeline` is a English model originally trained by utahnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724148934050.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/boolq_t5_base_seed_1_pipeline_en_5.4.2_3.0_1724148934050.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("boolq_t5_base_seed_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("boolq_t5_base_seed_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|boolq_t5_base_seed_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|963.9 MB| + +## References + +https://huggingface.co/utahnlp/boolq_t5-base_seed-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_en.md new file mode 100644 index 00000000000000..1288660a79dd1a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_modelv3 T5Transformer from Audino +author: John Snow Labs +name: burmese_awesome_modelv3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_modelv3` is a English model originally trained by Audino. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_modelv3_en_5.4.2_3.0_1724150239176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_modelv3_en_5.4.2_3.0_1724150239176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_modelv3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_modelv3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_modelv3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|975.5 MB| + +## References + +https://huggingface.co/Audino/my-awesome-modelv3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_pipeline_en.md new file mode 100644 index 00000000000000..d82b67f995f23f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_modelv3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_modelv3_pipeline pipeline T5Transformer from Audino +author: John Snow Labs +name: burmese_awesome_modelv3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_modelv3_pipeline` is a English model originally trained by Audino. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_modelv3_pipeline_en_5.4.2_3.0_1724150294069.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_modelv3_pipeline_en_5.4.2_3.0_1724150294069.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_modelv3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_modelv3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_modelv3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|975.5 MB| + +## References + +https://huggingface.co/Audino/my-awesome-modelv3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_en.md new file mode 100644 index 00000000000000..0d737cfaae2338 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_pacman1066work T5Transformer from pacman1066work +author: John Snow Labs +name: burmese_awesome_opus_books_model_pacman1066work +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_pacman1066work` is a English model originally trained by pacman1066work. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pacman1066work_en_5.4.2_3.0_1724113316634.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pacman1066work_en_5.4.2_3.0_1724113316634.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_pacman1066work","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_awesome_opus_books_model_pacman1066work", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_pacman1066work| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|331.4 MB| + +## References + +https://huggingface.co/pacman1066work/my_awesome_opus_books_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_pipeline_en.md new file mode 100644 index 00000000000000..6158ec4dabef63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_awesome_opus_books_model_pacman1066work_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_awesome_opus_books_model_pacman1066work_pipeline pipeline T5Transformer from pacman1066work +author: John Snow Labs +name: burmese_awesome_opus_books_model_pacman1066work_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_awesome_opus_books_model_pacman1066work_pipeline` is a English model originally trained by pacman1066work. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pacman1066work_pipeline_en_5.4.2_3.0_1724113339207.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_awesome_opus_books_model_pacman1066work_pipeline_en_5.4.2_3.0_1724113339207.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_awesome_opus_books_model_pacman1066work_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_awesome_opus_books_model_pacman1066work_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_awesome_opus_books_model_pacman1066work_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|331.4 MB| + +## References + +https://huggingface.co/pacman1066work/my_awesome_opus_books_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_en.md new file mode 100644 index 00000000000000..c60ab3ed0edf95 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_model_koltunov_matthew T5Transformer from Koltunov-Matthew +author: John Snow Labs +name: burmese_model_koltunov_matthew +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_model_koltunov_matthew` is a English model originally trained by Koltunov-Matthew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_model_koltunov_matthew_en_5.4.2_3.0_1724116026559.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_model_koltunov_matthew_en_5.4.2_3.0_1724116026559.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_model_koltunov_matthew","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_model_koltunov_matthew", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_model_koltunov_matthew| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/Koltunov-Matthew/my_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_pipeline_en.md new file mode 100644 index 00000000000000..62913c0c11f876 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_model_koltunov_matthew_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_model_koltunov_matthew_pipeline pipeline T5Transformer from Koltunov-Matthew +author: John Snow Labs +name: burmese_model_koltunov_matthew_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_model_koltunov_matthew_pipeline` is a English model originally trained by Koltunov-Matthew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_model_koltunov_matthew_pipeline_en_5.4.2_3.0_1724116045866.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_model_koltunov_matthew_pipeline_en_5.4.2_3.0_1724116045866.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_model_koltunov_matthew_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_model_koltunov_matthew_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_model_koltunov_matthew_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|336.2 MB| + +## References + +https://huggingface.co/Koltunov-Matthew/my_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_en.md new file mode 100644 index 00000000000000..c409869fa5f720 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_samsum_model T5Transformer from shubh-kandpal +author: John Snow Labs +name: burmese_samsum_model +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_samsum_model` is a English model originally trained by shubh-kandpal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_samsum_model_en_5.4.2_3.0_1724143691837.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_samsum_model_en_5.4.2_3.0_1724143691837.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_samsum_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_samsum_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_samsum_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|327.3 MB| + +## References + +https://huggingface.co/shubh-kandpal/my_samsum_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_pipeline_en.md new file mode 100644 index 00000000000000..bb114b916db158 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_samsum_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_samsum_model_pipeline pipeline T5Transformer from shubh-kandpal +author: John Snow Labs +name: burmese_samsum_model_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_samsum_model_pipeline` is a English model originally trained by shubh-kandpal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_samsum_model_pipeline_en_5.4.2_3.0_1724143712501.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_samsum_model_pipeline_en_5.4.2_3.0_1724143712501.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_samsum_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_samsum_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_samsum_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|327.3 MB| + +## References + +https://huggingface.co/shubh-kandpal/my_samsum_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_en.md new file mode 100644 index 00000000000000..8f9c39447aa342 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English burmese_t5_strange_answering T5Transformer from nosnic +author: John Snow Labs +name: burmese_t5_strange_answering +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_t5_strange_answering` is a English model originally trained by nosnic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_t5_strange_answering_en_5.4.2_3.0_1724141047866.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_t5_strange_answering_en_5.4.2_3.0_1724141047866.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("burmese_t5_strange_answering","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("burmese_t5_strange_answering", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_t5_strange_answering| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/nosnic/my_t5_strange_answering \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_pipeline_en.md new file mode 100644 index 00000000000000..839b7789bc5eba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-burmese_t5_strange_answering_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English burmese_t5_strange_answering_pipeline pipeline T5Transformer from nosnic +author: John Snow Labs +name: burmese_t5_strange_answering_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`burmese_t5_strange_answering_pipeline` is a English model originally trained by nosnic. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/burmese_t5_strange_answering_pipeline_en_5.4.2_3.0_1724141093990.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/burmese_t5_strange_answering_pipeline_en_5.4.2_3.0_1724141093990.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("burmese_t5_strange_answering_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("burmese_t5_strange_answering_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|burmese_t5_strange_answering_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/nosnic/my_t5_strange_answering + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_en.md b/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_en.md new file mode 100644 index 00000000000000..1f584200e9c9fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English catt5_solr_finetunned_complet2 T5Transformer from oooriii +author: John Snow Labs +name: catt5_solr_finetunned_complet2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`catt5_solr_finetunned_complet2` is a English model originally trained by oooriii. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned_complet2_en_5.4.2_3.0_1724126716785.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned_complet2_en_5.4.2_3.0_1724126716785.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("catt5_solr_finetunned_complet2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("catt5_solr_finetunned_complet2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|catt5_solr_finetunned_complet2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|918.9 MB| + +## References + +https://huggingface.co/oooriii/catt5-solr-finetunned_complet2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_pipeline_en.md new file mode 100644 index 00000000000000..5faab8e8012b86 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-catt5_solr_finetunned_complet2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English catt5_solr_finetunned_complet2_pipeline pipeline T5Transformer from oooriii +author: John Snow Labs +name: catt5_solr_finetunned_complet2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`catt5_solr_finetunned_complet2_pipeline` is a English model originally trained by oooriii. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned_complet2_pipeline_en_5.4.2_3.0_1724126783698.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/catt5_solr_finetunned_complet2_pipeline_en_5.4.2_3.0_1724126783698.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("catt5_solr_finetunned_complet2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("catt5_solr_finetunned_complet2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|catt5_solr_finetunned_complet2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|918.9 MB| + +## References + +https://huggingface.co/oooriii/catt5-solr-finetunned_complet2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_en.md b/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_en.md new file mode 100644 index 00000000000000..ae018357c4cf1f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cnn_aligned_smallt5_cont2 T5Transformer from paulh27 +author: John Snow Labs +name: cnn_aligned_smallt5_cont2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_aligned_smallt5_cont2` is a English model originally trained by paulh27. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_aligned_smallt5_cont2_en_5.4.2_3.0_1724129497925.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_aligned_smallt5_cont2_en_5.4.2_3.0_1724129497925.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cnn_aligned_smallt5_cont2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cnn_aligned_smallt5_cont2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_aligned_smallt5_cont2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/paulh27/cnn_aligned_smallT5_cont2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_pipeline_en.md new file mode 100644 index 00000000000000..a6e21f7240b9db --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cnn_aligned_smallt5_cont2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cnn_aligned_smallt5_cont2_pipeline pipeline T5Transformer from paulh27 +author: John Snow Labs +name: cnn_aligned_smallt5_cont2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_aligned_smallt5_cont2_pipeline` is a English model originally trained by paulh27. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_aligned_smallt5_cont2_pipeline_en_5.4.2_3.0_1724129514686.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_aligned_smallt5_cont2_pipeline_en_5.4.2_3.0_1724129514686.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cnn_aligned_smallt5_cont2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cnn_aligned_smallt5_cont2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_aligned_smallt5_cont2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.1 MB| + +## References + +https://huggingface.co/paulh27/cnn_aligned_smallT5_cont2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md b/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md new file mode 100644 index 00000000000000..8a5d728c0a0913 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6 T5Transformer from KingKazma +author: John Snow Labs +name: cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en_5.4.2_3.0_1724118437877.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_en_5.4.2_3.0_1724118437877.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|301.9 MB| + +## References + +https://huggingface.co/KingKazma/cnn_dailymail_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md new file mode 100644 index 00000000000000..42b54c3b7545a3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline pipeline T5Transformer from KingKazma +author: John Snow Labs +name: cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline` is a English model originally trained by KingKazma. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en_5.4.2_3.0_1724118465835.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline_en_5.4.2_3.0_1724118465835.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cnn_dailymail_t5_small_fine_tuning_500_10_3000_6_e_1_s6789_v4_l6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|301.9 MB| + +## References + +https://huggingface.co/KingKazma/cnn_dailymail_t5-small_fine_tuning_500_10_3000_6_e-1_s6789_v4_l6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_en.md new file mode 100644 index 00000000000000..a23d4a5eb85973 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs341_camera_coqe_unicoqe_t5small T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs341_camera_coqe_unicoqe_t5small +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs341_camera_coqe_unicoqe_t5small` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs341_camera_coqe_unicoqe_t5small_en_5.4.2_3.0_1724142981566.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs341_camera_coqe_unicoqe_t5small_en_5.4.2_3.0_1724142981566.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs341_camera_coqe_unicoqe_t5small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs341_camera_coqe_unicoqe_t5small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs341_camera_coqe_unicoqe_t5small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|343.1 MB| + +## References + +https://huggingface.co/ThuyNT03/CS341_Camera-COQE_UniCOQE_t5small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_pipeline_en.md new file mode 100644 index 00000000000000..6875c4ac847ab1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs341_camera_coqe_unicoqe_t5small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs341_camera_coqe_unicoqe_t5small_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs341_camera_coqe_unicoqe_t5small_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs341_camera_coqe_unicoqe_t5small_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs341_camera_coqe_unicoqe_t5small_pipeline_en_5.4.2_3.0_1724143000153.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs341_camera_coqe_unicoqe_t5small_pipeline_en_5.4.2_3.0_1724143000153.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs341_camera_coqe_unicoqe_t5small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs341_camera_coqe_unicoqe_t5small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs341_camera_coqe_unicoqe_t5small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|343.1 MB| + +## References + +https://huggingface.co/ThuyNT03/CS341_Camera-COQE_UniCOQE_t5small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_en.md new file mode 100644 index 00000000000000..5e39af9b628225 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_aspol_check_stable1 T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_aspol_check_stable1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_aspol_check_stable1` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_check_stable1_en_5.4.2_3.0_1724120171264.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_check_stable1_en_5.4.2_3.0_1724120171264.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_aspol_check_stable1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_prompting5_aspol_check_stable1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_aspol_check_stable1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_ASPOL_check_stable1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline_en.md new file mode 100644 index 00000000000000..b0e15ea0f32500 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline_en_5.4.2_3.0_1724120349491.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline_en_5.4.2_3.0_1724120349491.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_prompting5_aspol_check_stable1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/CS505_COQE_viT5_Prompting5_ASPOL_check_stable1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_en.md new file mode 100644 index 00000000000000..62502c4e332bb3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_aspol_v1_h0 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_aspol_v1_h0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_aspol_v1_h0` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aspol_v1_h0_en_5.4.2_3.0_1724119402302.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aspol_v1_h0_en_5.4.2_3.0_1724119402302.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_aspol_v1_h0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_aspol_v1_h0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_aspol_v1_h0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_ASPOL_v1_h0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline_en.md new file mode 100644 index 00000000000000..547675e156035c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline_en_5.4.2_3.0_1724119618789.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline_en_5.4.2_3.0_1724119618789.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_aspol_v1_h0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_ASPOL_v1_h0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_en.md new file mode 100644 index 00000000000000..6ba78b3a195069 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_saopl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_saopl_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_saopl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_en_5.4.2_3.0_1724149432196.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_en_5.4.2_3.0_1724149432196.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_saopl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_saopl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_saopl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAOPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline_en.md new file mode 100644 index 00000000000000..db07c7919b9040 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline_en_5.4.2_3.0_1724149617761.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline_en_5.4.2_3.0_1724149617761.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_saopl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAOPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_en.md new file mode 100644 index 00000000000000..14ea92249d7b8c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_sapol_v1_h1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_sapol_v1_h1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_sapol_v1_h1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_sapol_v1_h1_en_5.4.2_3.0_1724134778764.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_sapol_v1_h1_en_5.4.2_3.0_1724134778764.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_sapol_v1_h1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction0_sapol_v1_h1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_sapol_v1_h1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAPOL_v1_h1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline_en.md new file mode 100644 index 00000000000000..ed37f05e80ff39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline_en_5.4.2_3.0_1724134954653.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline_en_5.4.2_3.0_1724134954653.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction0_sapol_v1_h1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction0_SAPOL_v1_h1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction4_opasl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction4_opasl_v1_en.md new file mode 100644 index 00000000000000..818f531d26cde5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_total_instruction4_opasl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_total_instruction4_opasl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_total_instruction4_opasl_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_total_instruction4_opasl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction4_opasl_v1_en_5.4.2_3.0_1724142781424.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_total_instruction4_opasl_v1_en_5.4.2_3.0_1724142781424.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction4_opasl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_total_instruction4_opasl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_total_instruction4_opasl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_total_Instruction4_OPASL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_en.md new file mode 100644 index 00000000000000..fd5e0ce2453d85 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_osapl T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_osapl +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_osapl` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_osapl_en_5.4.2_3.0_1724147751078.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_osapl_en_5.4.2_3.0_1724147751078.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_osapl","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction0_osapl", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_osapl| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_OSAPL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_pipeline_en.md new file mode 100644 index 00000000000000..91581afbafe971 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction0_osapl_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction0_osapl_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction0_osapl_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction0_osapl_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_osapl_pipeline_en_5.4.2_3.0_1724147959084.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction0_osapl_pipeline_en_5.4.2_3.0_1724147959084.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction0_osapl_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction0_osapl_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction0_osapl_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction0_OSAPL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_en.md new file mode 100644 index 00000000000000..626281d34454e0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction4_paosl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction4_paosl_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction4_paosl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_paosl_v1_en_5.4.2_3.0_1724145715168.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_paosl_v1_en_5.4.2_3.0_1724145715168.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction4_paosl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instruction4_paosl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction4_paosl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction4_PAOSL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline_en.md new file mode 100644 index 00000000000000..8a3c88c42140ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline_en_5.4.2_3.0_1724145899358.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline_en_5.4.2_3.0_1724145899358.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instruction4_paosl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_Instruction4_PAOSL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_en.md new file mode 100644 index 00000000000000..c16f440a821507 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn0_ospal_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn0_ospal_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn0_ospal_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_ospal_v1_en_5.4.2_3.0_1724145984636.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_ospal_v1_en_5.4.2_3.0_1724145984636.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn0_ospal_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn0_ospal_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn0_ospal_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN0_OSPAL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline_en.md new file mode 100644 index 00000000000000..b7960bc2569665 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline_en_5.4.2_3.0_1724146167176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline_en_5.4.2_3.0_1724146167176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn0_ospal_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN0_OSPAL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_en.md new file mode 100644 index 00000000000000..98deef9f77003a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_soapl_v1 T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_soapl_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_soapl_v1` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_soapl_v1_en_5.4.2_3.0_1724130514659.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_soapl_v1_en_5.4.2_3.0_1724130514659.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_soapl_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("cs505_coqe_vit5_train_instructionn4_soapl_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_soapl_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_SOAPL_v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline_en.md new file mode 100644 index 00000000000000..77d129cae6c42b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline pipeline T5Transformer from ThuyNT +author: John Snow Labs +name: cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline` is a English model originally trained by ThuyNT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline_en_5.4.2_3.0_1724130692830.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline_en_5.4.2_3.0_1724130692830.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|cs505_coqe_vit5_train_instructionn4_soapl_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT/CS505_COQE_viT5_train_InstructionN4_SOAPL_v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_en.md b/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_en.md new file mode 100644 index 00000000000000..67633bee43dc91 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English deed_summarization_mt5_version_1 T5Transformer from Hasanur525 +author: John Snow Labs +name: deed_summarization_mt5_version_1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`deed_summarization_mt5_version_1` is a English model originally trained by Hasanur525. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/deed_summarization_mt5_version_1_en_5.4.2_3.0_1724148308023.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/deed_summarization_mt5_version_1_en_5.4.2_3.0_1724148308023.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("deed_summarization_mt5_version_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("deed_summarization_mt5_version_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|deed_summarization_mt5_version_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/Hasanur525/deed_summarization_mt5_version_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_pipeline_en.md new file mode 100644 index 00000000000000..6ca265e1e82539 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-deed_summarization_mt5_version_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English deed_summarization_mt5_version_1_pipeline pipeline T5Transformer from Hasanur525 +author: John Snow Labs +name: deed_summarization_mt5_version_1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`deed_summarization_mt5_version_1_pipeline` is a English model originally trained by Hasanur525. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/deed_summarization_mt5_version_1_pipeline_en_5.4.2_3.0_1724148554532.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/deed_summarization_mt5_version_1_pipeline_en_5.4.2_3.0_1724148554532.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("deed_summarization_mt5_version_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("deed_summarization_mt5_version_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|deed_summarization_mt5_version_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/Hasanur525/deed_summarization_mt5_version_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_en.md new file mode 100644 index 00000000000000..4cc56acbc39895 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_010099_8 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_010099_8 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_010099_8` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_8_en_5.4.2_3.0_1724124388028.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_8_en_5.4.2_3.0_1724124388028.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_010099_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_010099_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_010099_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-010099_8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_pipeline_en.md new file mode 100644 index 00000000000000..55471b25517767 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_010099_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_010099_8_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_010099_8_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_010099_8_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_8_pipeline_en_5.4.2_3.0_1724124579181.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_010099_8_pipeline_en_5.4.2_3.0_1724124579181.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_010099_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_010099_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_010099_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-010099_8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_en.md new file mode 100644 index 00000000000000..635f232a29a8ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_0_4_2 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_4_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_4_2` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_2_en_5.4.2_3.0_1724137932537.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_2_en_5.4.2_3.0_1724137932537.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_0_4_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_0_4_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_4_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.4-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_pipeline_en.md new file mode 100644 index 00000000000000..0a43bd9e1d8e25 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_0_4_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_0_4_2_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_0_4_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_0_4_2_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_2_pipeline_en_5.4.2_3.0_1724138119067.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_0_4_2_pipeline_en_5.4.2_3.0_1724138119067.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_0_4_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_0_4_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_0_4_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-0.4-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_en.md new file mode 100644 index 00000000000000..c57296bd35e5ab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_b10 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b10 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b10` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b10_en_5.4.2_3.0_1724120554700.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b10_en_5.4.2_3.0_1724120554700.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_b10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_b10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_pipeline_en.md new file mode 100644 index 00000000000000..78f7bc9ef1ddf1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_b10_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b10_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b10_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b10_pipeline_en_5.4.2_3.0_1724120749045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b10_pipeline_en_5.4.2_3.0_1724120749045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_b10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_b10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_en.md new file mode 100644 index 00000000000000..0c21aac2b7867d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English distilled_mt5_small_b1_5 T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b1_5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b1_5` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b1_5_en_5.4.2_3.0_1724124884077.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b1_5_en_5.4.2_3.0_1724124884077.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("distilled_mt5_small_b1_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("distilled_mt5_small_b1_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b1_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b1.5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_pipeline_en.md new file mode 100644 index 00000000000000..557e5a113a45aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-distilled_mt5_small_b1_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English distilled_mt5_small_b1_5_pipeline pipeline T5Transformer from Lvxue +author: John Snow Labs +name: distilled_mt5_small_b1_5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`distilled_mt5_small_b1_5_pipeline` is a English model originally trained by Lvxue. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b1_5_pipeline_en_5.4.2_3.0_1724125075410.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/distilled_mt5_small_b1_5_pipeline_en_5.4.2_3.0_1724125075410.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("distilled_mt5_small_b1_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("distilled_mt5_small_b1_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|distilled_mt5_small_b1_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Lvxue/distilled-mt5-small-b1.5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_en.md b/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_en.md new file mode 100644 index 00000000000000..27daa1b02a6d39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English docut5_base_sindhi T5Transformer from totem37 +author: John Snow Labs +name: docut5_base_sindhi +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`docut5_base_sindhi` is a English model originally trained by totem37. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/docut5_base_sindhi_en_5.4.2_3.0_1724125525221.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/docut5_base_sindhi_en_5.4.2_3.0_1724125525221.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("docut5_base_sindhi","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("docut5_base_sindhi", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|docut5_base_sindhi| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/totem37/DocuT5-Base-SD \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_pipeline_en.md new file mode 100644 index 00000000000000..163c545fbf99a7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-docut5_base_sindhi_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English docut5_base_sindhi_pipeline pipeline T5Transformer from totem37 +author: John Snow Labs +name: docut5_base_sindhi_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`docut5_base_sindhi_pipeline` is a English model originally trained by totem37. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/docut5_base_sindhi_pipeline_en_5.4.2_3.0_1724125573113.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/docut5_base_sindhi_pipeline_en_5.4.2_3.0_1724125573113.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("docut5_base_sindhi_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("docut5_base_sindhi_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|docut5_base_sindhi_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/totem37/DocuT5-Base-SD + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-dug_t5large_0_1_en.md b/docs/_posts/ahmedlone127/2024-08-20-dug_t5large_0_1_en.md new file mode 100644 index 00000000000000..88d5f9b2354067 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-dug_t5large_0_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dug_t5large_0_1 T5Transformer from OpenCUI +author: John Snow Labs +name: dug_t5large_0_1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dug_t5large_0_1` is a English model originally trained by OpenCUI. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dug_t5large_0_1_en_5.4.2_3.0_1724112138996.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dug_t5large_0_1_en_5.4.2_3.0_1724112138996.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dug_t5large_0_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dug_t5large_0_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dug_t5large_0_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/OpenCUI/dug-t5large-0.1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_en.md b/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_en.md new file mode 100644 index 00000000000000..0f691c6bf63ade --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English dutch_ge_25_6b_3ep_northern_sami T5Transformer from Bistolero +author: John Snow Labs +name: dutch_ge_25_6b_3ep_northern_sami +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dutch_ge_25_6b_3ep_northern_sami` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dutch_ge_25_6b_3ep_northern_sami_en_5.4.2_3.0_1724141257045.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dutch_ge_25_6b_3ep_northern_sami_en_5.4.2_3.0_1724141257045.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("dutch_ge_25_6b_3ep_northern_sami","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("dutch_ge_25_6b_3ep_northern_sami", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dutch_ge_25_6b_3ep_northern_sami| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/nl_ge_25_6b_3ep_se \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_pipeline_en.md new file mode 100644 index 00000000000000..b30bde1f891f32 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-dutch_ge_25_6b_3ep_northern_sami_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English dutch_ge_25_6b_3ep_northern_sami_pipeline pipeline T5Transformer from Bistolero +author: John Snow Labs +name: dutch_ge_25_6b_3ep_northern_sami_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`dutch_ge_25_6b_3ep_northern_sami_pipeline` is a English model originally trained by Bistolero. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/dutch_ge_25_6b_3ep_northern_sami_pipeline_en_5.4.2_3.0_1724141417906.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/dutch_ge_25_6b_3ep_northern_sami_pipeline_en_5.4.2_3.0_1724141417906.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("dutch_ge_25_6b_3ep_northern_sami_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("dutch_ge_25_6b_3ep_northern_sami_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|dutch_ge_25_6b_3ep_northern_sami_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/Bistolero/nl_ge_25_6b_3ep_se + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_en.md b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_en.md new file mode 100644 index 00000000000000..de29c6e588c6ef --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_vietnamese_envit5_base_news_train T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_envit5_base_news_train +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_envit5_base_news_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_base_news_train_en_5.4.2_3.0_1724125890257.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_base_news_train_en_5.4.2_3.0_1724125890257.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_vietnamese_envit5_base_news_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_vietnamese_envit5_base_news_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_envit5_base_news_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_envit5-base_news_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_pipeline_en.md new file mode 100644 index 00000000000000..0710c2e2670230 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_envit5_base_news_train_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_vietnamese_envit5_base_news_train_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_envit5_base_news_train_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_envit5_base_news_train_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_base_news_train_pipeline_en_5.4.2_3.0_1724125966144.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_envit5_base_news_train_pipeline_en_5.4.2_3.0_1724125966144.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_vietnamese_envit5_base_news_train_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_vietnamese_envit5_base_news_train_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_envit5_base_news_train_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_envit5-base_news_train + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_en.md b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_en.md new file mode 100644 index 00000000000000..64e5a0a6453bcc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English english_vietnamese_mt5_base_news_train T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_mt5_base_news_train +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_mt5_base_news_train` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_news_train_en_5.4.2_3.0_1724140265049.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_news_train_en_5.4.2_3.0_1724140265049.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("english_vietnamese_mt5_base_news_train","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("english_vietnamese_mt5_base_news_train", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_mt5_base_news_train| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_mt5-base_news_train \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_pipeline_en.md new file mode 100644 index 00000000000000..c6ce60ab4235b9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-english_vietnamese_mt5_base_news_train_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English english_vietnamese_mt5_base_news_train_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: english_vietnamese_mt5_base_news_train_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`english_vietnamese_mt5_base_news_train_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_news_train_pipeline_en_5.4.2_3.0_1724140533160.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/english_vietnamese_mt5_base_news_train_pipeline_en_5.4.2_3.0_1724140533160.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("english_vietnamese_mt5_base_news_train_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("english_vietnamese_mt5_base_news_train_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|english_vietnamese_mt5_base_news_train_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.1 GB| + +## References + +https://huggingface.co/hungphongtrn/en_vi_mt5-base_news_train + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-engtofil_en.md b/docs/_posts/ahmedlone127/2024-08-20-engtofil_en.md new file mode 100644 index 00000000000000..07df02d5550bf0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-engtofil_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English engtofil T5Transformer from roval15 +author: John Snow Labs +name: engtofil +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`engtofil` is a English model originally trained by roval15. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/engtofil_en_5.4.2_3.0_1724120304941.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/engtofil_en_5.4.2_3.0_1724120304941.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("engtofil","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("engtofil", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|engtofil| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/roval15/EngToFil \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-engtofil_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-engtofil_pipeline_en.md new file mode 100644 index 00000000000000..cd8245d1e6bde1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-engtofil_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English engtofil_pipeline pipeline T5Transformer from roval15 +author: John Snow Labs +name: engtofil_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`engtofil_pipeline` is a English model originally trained by roval15. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/engtofil_pipeline_en_5.4.2_3.0_1724120354172.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/engtofil_pipeline_en_5.4.2_3.0_1724120354172.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("engtofil_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("engtofil_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|engtofil_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/roval15/EngToFil + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_en.md b/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_en.md new file mode 100644 index 00000000000000..ef56261ae9c0a9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English finetune_keyword_t5_tiny_standard_bahasa_cased T5Transformer from mesolitica +author: John Snow Labs +name: finetune_keyword_t5_tiny_standard_bahasa_cased +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_keyword_t5_tiny_standard_bahasa_cased` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_tiny_standard_bahasa_cased_en_5.4.2_3.0_1724123712596.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_tiny_standard_bahasa_cased_en_5.4.2_3.0_1724123712596.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("finetune_keyword_t5_tiny_standard_bahasa_cased","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("finetune_keyword_t5_tiny_standard_bahasa_cased", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_keyword_t5_tiny_standard_bahasa_cased| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|222.9 MB| + +## References + +https://huggingface.co/mesolitica/finetune-keyword-t5-tiny-standard-bahasa-cased \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline_en.md new file mode 100644 index 00000000000000..3e92215cbc5ffc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline pipeline T5Transformer from mesolitica +author: John Snow Labs +name: finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline` is a English model originally trained by mesolitica. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724123723649.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline_en_5.4.2_3.0_1724123723649.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|finetune_keyword_t5_tiny_standard_bahasa_cased_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|222.9 MB| + +## References + +https://huggingface.co/mesolitica/finetune-keyword-t5-tiny-standard-bahasa-cased + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_en.md new file mode 100644 index 00000000000000..5bbcce311c93f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_large_passage_evidence T5Transformer from erbacher +author: John Snow Labs +name: flan_large_passage_evidence +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_large_passage_evidence` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_large_passage_evidence_en_5.4.2_3.0_1724123849374.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_large_passage_evidence_en_5.4.2_3.0_1724123849374.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_large_passage_evidence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_large_passage_evidence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_large_passage_evidence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/erbacher/flan-large-passage-evidence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_pipeline_en.md new file mode 100644 index 00000000000000..5db261322953e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_large_passage_evidence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_large_passage_evidence_pipeline pipeline T5Transformer from erbacher +author: John Snow Labs +name: flan_large_passage_evidence_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_large_passage_evidence_pipeline` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_large_passage_evidence_pipeline_en_5.4.2_3.0_1724123986943.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_large_passage_evidence_pipeline_en_5.4.2_3.0_1724123986943.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_large_passage_evidence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_large_passage_evidence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_large_passage_evidence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/erbacher/flan-large-passage-evidence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_en.md new file mode 100644 index 00000000000000..ef374de464be3a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_1_1_cnndm T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_1_cnndm +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_1_cnndm` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_1_cnndm_en_5.4.2_3.0_1724122912937.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_1_cnndm_en_5.4.2_3.0_1724122912937.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_1_1_cnndm","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_1_1_cnndm", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_1_cnndm| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|400.4 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-1-cnndm \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_pipeline_en.md new file mode 100644 index 00000000000000..f3fdb9c44b5beb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_1_1_cnndm_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_1_1_cnndm_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_1_1_cnndm_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_1_1_cnndm_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_1_cnndm_pipeline_en_5.4.2_3.0_1724122931882.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_1_1_cnndm_pipeline_en_5.4.2_3.0_1724122931882.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_1_1_cnndm_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_1_1_cnndm_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_1_1_cnndm_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|400.4 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-1-1-cnndm + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_en.md new file mode 100644 index 00000000000000..d8e4e1c7d46907 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_3_6_xsum T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_3_6_xsum +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_3_6_xsum` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_xsum_en_5.4.2_3.0_1724112640311.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_xsum_en_5.4.2_3.0_1724112640311.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_3_6_xsum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_3_6_xsum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_3_6_xsum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|860.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-3-6-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_pipeline_en.md new file mode 100644 index 00000000000000..34b4c0b8e406df --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_3_6_xsum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_3_6_xsum_pipeline pipeline T5Transformer from spacemanidol +author: John Snow Labs +name: flan_t5_base_3_6_xsum_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_3_6_xsum_pipeline` is a English model originally trained by spacemanidol. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_xsum_pipeline_en_5.4.2_3.0_1724112681600.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_3_6_xsum_pipeline_en_5.4.2_3.0_1724112681600.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_3_6_xsum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_3_6_xsum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_3_6_xsum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|860.1 MB| + +## References + +https://huggingface.co/spacemanidol/flan-t5-base-3-6-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_en.md new file mode 100644 index 00000000000000..5a6705aa59d3e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_analogy_conceptnet T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_base_analogy_conceptnet +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_analogy_conceptnet` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_conceptnet_en_5.4.2_3.0_1724125579611.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_conceptnet_en_5.4.2_3.0_1724125579611.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_analogy_conceptnet","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_analogy_conceptnet", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_analogy_conceptnet| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/flan-t5-base-analogy-conceptnet \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_pipeline_en.md new file mode 100644 index 00000000000000..469eaa292e0474 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_analogy_conceptnet_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_analogy_conceptnet_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: flan_t5_base_analogy_conceptnet_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_analogy_conceptnet_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_conceptnet_pipeline_en_5.4.2_3.0_1724125628269.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_analogy_conceptnet_pipeline_en_5.4.2_3.0_1724125628269.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_analogy_conceptnet_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_analogy_conceptnet_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_analogy_conceptnet_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/research-backup/flan-t5-base-analogy-conceptnet + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_en.md new file mode 100644 index 00000000000000..823b828adddfa4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_samsum_dutch_split_ep5 T5Transformer from yingzwang +author: John Snow Labs +name: flan_t5_base_samsum_dutch_split_ep5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_samsum_dutch_split_ep5` is a English model originally trained by yingzwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_dutch_split_ep5_en_5.4.2_3.0_1724116658115.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_dutch_split_ep5_en_5.4.2_3.0_1724116658115.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_samsum_dutch_split_ep5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_samsum_dutch_split_ep5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_samsum_dutch_split_ep5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yingzwang/flan-t5-base-samsum_nl_split_ep5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_pipeline_en.md new file mode 100644 index 00000000000000..947dc23c8d06db --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_samsum_dutch_split_ep5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_samsum_dutch_split_ep5_pipeline pipeline T5Transformer from yingzwang +author: John Snow Labs +name: flan_t5_base_samsum_dutch_split_ep5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_samsum_dutch_split_ep5_pipeline` is a English model originally trained by yingzwang. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_dutch_split_ep5_pipeline_en_5.4.2_3.0_1724116715108.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_samsum_dutch_split_ep5_pipeline_en_5.4.2_3.0_1724116715108.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_samsum_dutch_split_ep5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_samsum_dutch_split_ep5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_samsum_dutch_split_ep5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yingzwang/flan-t5-base-samsum_nl_split_ep5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_en.md new file mode 100644 index 00000000000000..bee519d005a49c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_base_sherif T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_sherif +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_sherif` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_sherif_en_5.4.2_3.0_1724113021624.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_sherif_en_5.4.2_3.0_1724113021624.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_base_sherif","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_base_sherif", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_sherif| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-sherif \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_pipeline_en.md new file mode 100644 index 00000000000000..bb948cbea770f8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_base_sherif_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_base_sherif_pipeline pipeline T5Transformer from sherif1311 +author: John Snow Labs +name: flan_t5_base_sherif_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_base_sherif_pipeline` is a English model originally trained by sherif1311. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_base_sherif_pipeline_en_5.4.2_3.0_1724113069290.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_base_sherif_pipeline_en_5.4.2_3.0_1724113069290.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_base_sherif_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_base_sherif_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_base_sherif_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sherif1311/flan-t5-base-sherif + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_en.md new file mode 100644 index 00000000000000..9629e50a849437 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_800_loss_ep50 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_800_loss_ep50 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_800_loss_ep50` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_loss_ep50_en_5.4.2_3.0_1724123896920.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_loss_ep50_en_5.4.2_3.0_1724123896920.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_800_loss_ep50","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_0_800_loss_ep50", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_800_loss_ep50| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_800-loss-ep50 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline_en.md new file mode 100644 index 00000000000000..15ffd20c62b3da --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline_en_5.4.2_3.0_1724124041325.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline_en_5.4.2_3.0_1724124041325.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_0_800_loss_ep50_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.0_800-loss-ep50 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_en.md new file mode 100644 index 00000000000000..1f086736761c43 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_fs0_01 T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_fs0_01 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_fs0_01` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_01_en_5.4.2_3.0_1724139741534.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_01_en_5.4.2_3.0_1724139741534.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_fs0_01","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_danish_multiwoz2_1_fs0_01", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_fs0_01| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_fs0.01 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline_en.md new file mode 100644 index 00000000000000..4f78bed8ee286f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline pipeline T5Transformer from Zekunli +author: John Snow Labs +name: flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline` is a English model originally trained by Zekunli. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline_en_5.4.2_3.0_1724139870964.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline_en_5.4.2_3.0_1724139870964.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_danish_multiwoz2_1_fs0_01_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Zekunli/flan-t5-large-da-multiwoz2.1_fs0.01 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_fold_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_fold_2_en.md new file mode 100644 index 00000000000000..f0f7f350410428 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_fold_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_fold_2 T5Transformer from research-dump +author: John Snow Labs +name: flan_t5_large_fold_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_fold_2` is a English model originally trained by research-dump. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_fold_2_en_5.4.2_3.0_1724145392277.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_fold_2_en_5.4.2_3.0_1724145392277.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_fold_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_fold_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_fold_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/research-dump/flan-t5-large_fold_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_en.md new file mode 100644 index 00000000000000..1ec0ce0c1afb0b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_large_medistill_hierarchical_ep20 T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_hierarchical_ep20 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_hierarchical_ep20` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_hierarchical_ep20_en_5.4.2_3.0_1724120952461.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_hierarchical_ep20_en_5.4.2_3.0_1724120952461.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_large_medistill_hierarchical_ep20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_large_medistill_hierarchical_ep20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_hierarchical_ep20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_hierarchical_ep20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_pipeline_en.md new file mode 100644 index 00000000000000..7a70fd4f165eb9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_large_medistill_hierarchical_ep20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_large_medistill_hierarchical_ep20_pipeline pipeline T5Transformer from Xiaolihai +author: John Snow Labs +name: flan_t5_large_medistill_hierarchical_ep20_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_large_medistill_hierarchical_ep20_pipeline` is a English model originally trained by Xiaolihai. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_hierarchical_ep20_pipeline_en_5.4.2_3.0_1724121097453.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_large_medistill_hierarchical_ep20_pipeline_en_5.4.2_3.0_1724121097453.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_large_medistill_hierarchical_ep20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_large_medistill_hierarchical_ep20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_large_medistill_hierarchical_ep20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/Xiaolihai/flan-t5-large_MeDistill_hierarchical_ep20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_en.md new file mode 100644 index 00000000000000..bb049d75899720 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_asap_t3_f4_prompt_adherence T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t3_f4_prompt_adherence +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t3_f4_prompt_adherence` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t3_f4_prompt_adherence_en_5.4.2_3.0_1724117901237.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t3_f4_prompt_adherence_en_5.4.2_3.0_1724117901237.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_asap_t3_f4_prompt_adherence","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_asap_t3_f4_prompt_adherence", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t3_f4_prompt_adherence| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t3_f4_prompt_adherence \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_pipeline_en.md new file mode 100644 index 00000000000000..26d8eccdfd248a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_asap_t3_f4_prompt_adherence_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_asap_t3_f4_prompt_adherence_pipeline pipeline T5Transformer from salbatarni +author: John Snow Labs +name: flan_t5_small_asap_t3_f4_prompt_adherence_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_asap_t3_f4_prompt_adherence_pipeline` is a English model originally trained by salbatarni. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t3_f4_prompt_adherence_pipeline_en_5.4.2_3.0_1724117918200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_asap_t3_f4_prompt_adherence_pipeline_en_5.4.2_3.0_1724117918200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_asap_t3_f4_prompt_adherence_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_asap_t3_f4_prompt_adherence_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_asap_t3_f4_prompt_adherence_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/salbatarni/flan-t5-small-asap_t3_f4_prompt_adherence + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_en.md new file mode 100644 index 00000000000000..fa3e16b8147466 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_comma_correction T5Transformer from pavlichenko +author: John Snow Labs +name: flan_t5_small_comma_correction +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_comma_correction` is a English model originally trained by pavlichenko. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_comma_correction_en_5.4.2_3.0_1724143137133.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_comma_correction_en_5.4.2_3.0_1724143137133.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_comma_correction","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_comma_correction", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_comma_correction| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/pavlichenko/flan-t5-small-comma-correction \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_pipeline_en.md new file mode 100644 index 00000000000000..e2ff84378b81b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_comma_correction_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_comma_correction_pipeline pipeline T5Transformer from pavlichenko +author: John Snow Labs +name: flan_t5_small_comma_correction_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_comma_correction_pipeline` is a English model originally trained by pavlichenko. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_comma_correction_pipeline_en_5.4.2_3.0_1724143153542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_comma_correction_pipeline_en_5.4.2_3.0_1724143153542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_comma_correction_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_comma_correction_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_comma_correction_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/pavlichenko/flan-t5-small-comma-correction + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_en.md new file mode 100644 index 00000000000000..f444c217ed3ba4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_6 T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_6 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_6` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_6_en_5.4.2_3.0_1724133316568.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_6_en_5.4.2_3.0_1724133316568.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_finetuned_coqa_v0_6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_pipeline_en.md new file mode 100644 index 00000000000000..688c05c257dbda --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_finetuned_coqa_v0_6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_finetuned_coqa_v0_6_pipeline pipeline T5Transformer from rooftopcoder +author: John Snow Labs +name: flan_t5_small_finetuned_coqa_v0_6_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_finetuned_coqa_v0_6_pipeline` is a English model originally trained by rooftopcoder. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_6_pipeline_en_5.4.2_3.0_1724133333278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_finetuned_coqa_v0_6_pipeline_en_5.4.2_3.0_1724133333278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_finetuned_coqa_v0_6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_finetuned_coqa_v0_6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/rooftopcoder/flan-t5-small-finetuned-coqa-V0.6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_en.md new file mode 100644 index 00000000000000..8feb4a74d69aba --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flan_t5_small_tweet_emotion T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_emotion +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_emotion` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_emotion_en_5.4.2_3.0_1724117715615.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_emotion_en_5.4.2_3.0_1724117715615.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flan_t5_small_tweet_emotion","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flan_t5_small_tweet_emotion", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_emotion| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-emotion \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_pipeline_en.md new file mode 100644 index 00000000000000..6203ed13009eb8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flan_t5_small_tweet_emotion_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flan_t5_small_tweet_emotion_pipeline pipeline T5Transformer from cardiffnlp +author: John Snow Labs +name: flan_t5_small_tweet_emotion_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flan_t5_small_tweet_emotion_pipeline` is a English model originally trained by cardiffnlp. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_emotion_pipeline_en_5.4.2_3.0_1724117734657.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flan_t5_small_tweet_emotion_pipeline_en_5.4.2_3.0_1724117734657.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flan_t5_small_tweet_emotion_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flan_t5_small_tweet_emotion_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flan_t5_small_tweet_emotion_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.8 MB| + +## References + +https://huggingface.co/cardiffnlp/flan-t5-small-tweet-emotion + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_en.md b/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_en.md new file mode 100644 index 00000000000000..b42e9b74bb6bd7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English flanwithoutpassage T5Transformer from erbacher +author: John Snow Labs +name: flanwithoutpassage +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flanwithoutpassage` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flanwithoutpassage_en_5.4.2_3.0_1724147850286.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flanwithoutpassage_en_5.4.2_3.0_1724147850286.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("flanwithoutpassage","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("flanwithoutpassage", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flanwithoutpassage| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/erbacher/flanwithoutpassage \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_pipeline_en.md new file mode 100644 index 00000000000000..464ca2f028ce39 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-flanwithoutpassage_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English flanwithoutpassage_pipeline pipeline T5Transformer from erbacher +author: John Snow Labs +name: flanwithoutpassage_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`flanwithoutpassage_pipeline` is a English model originally trained by erbacher. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/flanwithoutpassage_pipeline_en_5.4.2_3.0_1724148012017.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/flanwithoutpassage_pipeline_en_5.4.2_3.0_1724148012017.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("flanwithoutpassage_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("flanwithoutpassage_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|flanwithoutpassage_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/erbacher/flanwithoutpassage + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_en.md b/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_en.md new file mode 100644 index 00000000000000..9792e37dcf4036 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English frame_semantic_transformer_google_t5_efficient_tiny T5Transformer from striebel +author: John Snow Labs +name: frame_semantic_transformer_google_t5_efficient_tiny +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`frame_semantic_transformer_google_t5_efficient_tiny` is a English model originally trained by striebel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/frame_semantic_transformer_google_t5_efficient_tiny_en_5.4.2_3.0_1724146607988.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/frame_semantic_transformer_google_t5_efficient_tiny_en_5.4.2_3.0_1724146607988.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("frame_semantic_transformer_google_t5_efficient_tiny","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("frame_semantic_transformer_google_t5_efficient_tiny", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|frame_semantic_transformer_google_t5_efficient_tiny| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|116.1 MB| + +## References + +https://huggingface.co/striebel/frame-semantic-transformer-google-t5-efficient-tiny \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_pipeline_en.md new file mode 100644 index 00000000000000..18ff30c3381f30 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-frame_semantic_transformer_google_t5_efficient_tiny_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English frame_semantic_transformer_google_t5_efficient_tiny_pipeline pipeline T5Transformer from striebel +author: John Snow Labs +name: frame_semantic_transformer_google_t5_efficient_tiny_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`frame_semantic_transformer_google_t5_efficient_tiny_pipeline` is a English model originally trained by striebel. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/frame_semantic_transformer_google_t5_efficient_tiny_pipeline_en_5.4.2_3.0_1724146615219.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/frame_semantic_transformer_google_t5_efficient_tiny_pipeline_en_5.4.2_3.0_1724146615219.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("frame_semantic_transformer_google_t5_efficient_tiny_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("frame_semantic_transformer_google_t5_efficient_tiny_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|frame_semantic_transformer_google_t5_efficient_tiny_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|116.1 MB| + +## References + +https://huggingface.co/striebel/frame-semantic-transformer-google-t5-efficient-tiny + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_en.md new file mode 100644 index 00000000000000..e47e3818033b2f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_aeroqa_1hop_aviationcorpusandaviationtriples_20 T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_aviationcorpusandaviationtriples_20 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_aviationcorpusandaviationtriples_20` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_en_5.4.2_3.0_1724142678009.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_en_5.4.2_3.0_1724142678009.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_aeroqa_1hop_aviationcorpusandaviationtriples_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_aeroqa_1hop_aviationcorpusandaviationtriples_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_aviationcorpusandaviationtriples_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_aviationcorpusAndaviationtriples_20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline_en.md new file mode 100644 index 00000000000000..e45a42a5447021 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline_en_5.4.2_3.0_1724142812067.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline_en_5.4.2_3.0_1724142812067.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_aviationcorpusandaviationtriples_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_aviationcorpusAndaviationtriples_20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_en.md new file mode 100644 index 00000000000000..05aa945c33efab --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_aeroqa_1hop_aviationtriples_unverb_20 T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_aviationtriples_unverb_20 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_aviationtriples_unverb_20` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationtriples_unverb_20_en_5.4.2_3.0_1724143494977.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationtriples_unverb_20_en_5.4.2_3.0_1724143494977.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_aeroqa_1hop_aviationtriples_unverb_20","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_aeroqa_1hop_aviationtriples_unverb_20", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_aviationtriples_unverb_20| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_aviationtriples_unverb_20 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline_en.md new file mode 100644 index 00000000000000..fafb023d0a55e3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline_en_5.4.2_3.0_1724143642329.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline_en_5.4.2_3.0_1724143642329.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_aeroqa_1hop_aviationtriples_unverb_20_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_aeroqa_1hop_aviationtriples_unverb_20 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_en.md new file mode 100644 index 00000000000000..acb1b2605c9cd7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English ft_metaqa_3hop_t5_largecheckpoint_117000 T5Transformer from sakharamg +author: John Snow Labs +name: ft_metaqa_3hop_t5_largecheckpoint_117000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_metaqa_3hop_t5_largecheckpoint_117000` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_metaqa_3hop_t5_largecheckpoint_117000_en_5.4.2_3.0_1724122459436.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_metaqa_3hop_t5_largecheckpoint_117000_en_5.4.2_3.0_1724122459436.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("ft_metaqa_3hop_t5_largecheckpoint_117000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("ft_metaqa_3hop_t5_largecheckpoint_117000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_metaqa_3hop_t5_largecheckpoint_117000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_metaqa_3hop_t5-largecheckpoint-117000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline_en.md new file mode 100644 index 00000000000000..1ba28dd0c81b83 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline pipeline T5Transformer from sakharamg +author: John Snow Labs +name: ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline` is a English model originally trained by sakharamg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline_en_5.4.2_3.0_1724122593253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline_en_5.4.2_3.0_1724122593253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|ft_metaqa_3hop_t5_largecheckpoint_117000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/sakharamg/FT_metaqa_3hop_t5-largecheckpoint-117000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-grammer_en.md b/docs/_posts/ahmedlone127/2024-08-20-grammer_en.md new file mode 100644 index 00000000000000..2410c814e7c421 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-grammer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English grammer T5Transformer from HARSHU550 +author: John Snow Labs +name: grammer +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`grammer` is a English model originally trained by HARSHU550. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/grammer_en_5.4.2_3.0_1724150556024.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/grammer_en_5.4.2_3.0_1724150556024.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("grammer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("grammer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|grammer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/HARSHU550/Grammer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-grammer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-grammer_pipeline_en.md new file mode 100644 index 00000000000000..0d9d51745ea5e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-grammer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English grammer_pipeline pipeline T5Transformer from HARSHU550 +author: John Snow Labs +name: grammer_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`grammer_pipeline` is a English model originally trained by HARSHU550. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/grammer_pipeline_en_5.4.2_3.0_1724150697483.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/grammer_pipeline_en_5.4.2_3.0_1724150697483.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("grammer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("grammer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|grammer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/HARSHU550/Grammer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_en.md b/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_en.md new file mode 100644 index 00000000000000..244beba8d2e35f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_lr_v3 T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_lr_v3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_lr_v3` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_lr_v3_en_5.4.2_3.0_1724133170046.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_lr_v3_en_5.4.2_3.0_1724133170046.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_lr_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("happy_transformer_t5_base_grammar_correction_lr_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_lr_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-lr-v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_pipeline_en.md new file mode 100644 index 00000000000000..e8f08ec68876a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-happy_transformer_t5_base_grammar_correction_lr_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English happy_transformer_t5_base_grammar_correction_lr_v3_pipeline pipeline T5Transformer from hafidikhsan +author: John Snow Labs +name: happy_transformer_t5_base_grammar_correction_lr_v3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`happy_transformer_t5_base_grammar_correction_lr_v3_pipeline` is a English model originally trained by hafidikhsan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_lr_v3_pipeline_en_5.4.2_3.0_1724133218557.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/happy_transformer_t5_base_grammar_correction_lr_v3_pipeline_en_5.4.2_3.0_1724133218557.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("happy_transformer_t5_base_grammar_correction_lr_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("happy_transformer_t5_base_grammar_correction_lr_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|happy_transformer_t5_base_grammar_correction_lr_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/hafidikhsan/happy-transformer-t5-base-grammar-correction-lr-v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_en.md b/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_en.md new file mode 100644 index 00000000000000..ea8a7032d232a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English inisw08_t5_mlm_adafactor_test T5Transformer from ugiugi +author: John Snow Labs +name: inisw08_t5_mlm_adafactor_test +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`inisw08_t5_mlm_adafactor_test` is a English model originally trained by ugiugi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/inisw08_t5_mlm_adafactor_test_en_5.4.2_3.0_1724124616380.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/inisw08_t5_mlm_adafactor_test_en_5.4.2_3.0_1724124616380.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("inisw08_t5_mlm_adafactor_test","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("inisw08_t5_mlm_adafactor_test", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|inisw08_t5_mlm_adafactor_test| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ugiugi/inisw08-T5-mlm-adafactor_test \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_pipeline_en.md new file mode 100644 index 00000000000000..33c6a87bfcc9b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-inisw08_t5_mlm_adafactor_test_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English inisw08_t5_mlm_adafactor_test_pipeline pipeline T5Transformer from ugiugi +author: John Snow Labs +name: inisw08_t5_mlm_adafactor_test_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`inisw08_t5_mlm_adafactor_test_pipeline` is a English model originally trained by ugiugi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/inisw08_t5_mlm_adafactor_test_pipeline_en_5.4.2_3.0_1724124662891.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/inisw08_t5_mlm_adafactor_test_pipeline_en_5.4.2_3.0_1724124662891.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("inisw08_t5_mlm_adafactor_test_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("inisw08_t5_mlm_adafactor_test_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|inisw08_t5_mlm_adafactor_test_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ugiugi/inisw08-T5-mlm-adafactor_test + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_en.md b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_en.md new file mode 100644 index 00000000000000..d823ae314dfac5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_aspol_v3 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_aspol_v3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_aspol_v3` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_aspol_v3_en_5.4.2_3.0_1724129005245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_aspol_v3_en_5.4.2_3.0_1724129005245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_aspol_v3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_aspol_v3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_aspol_v3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_ASPOL_v3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_pipeline_en.md new file mode 100644 index 00000000000000..58ef638ec73c0e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_aspol_v3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_aspol_v3_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_aspol_v3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_aspol_v3_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_aspol_v3_pipeline_en_5.4.2_3.0_1724129180548.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_aspol_v3_pipeline_en_5.4.2_3.0_1724129180548.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_aspol_v3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_aspol_v3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_aspol_v3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_ASPOL_v3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_en.md b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_en.md new file mode 100644 index 00000000000000..712e4f41dab66a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_spaol_v5 T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_spaol_v5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_spaol_v5` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v5_en_5.4.2_3.0_1724143730065.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v5_en_5.4.2_3.0_1724143730065.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_spaol_v5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kltn_coqe_vit5_total_spaol_v5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_spaol_v5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SPAOL_v5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_pipeline_en.md new file mode 100644 index 00000000000000..b49419a66adc4c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kltn_coqe_vit5_total_spaol_v5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kltn_coqe_vit5_total_spaol_v5_pipeline pipeline T5Transformer from ThuyNT03 +author: John Snow Labs +name: kltn_coqe_vit5_total_spaol_v5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kltn_coqe_vit5_total_spaol_v5_pipeline` is a English model originally trained by ThuyNT03. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v5_pipeline_en_5.4.2_3.0_1724143900134.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kltn_coqe_vit5_total_spaol_v5_pipeline_en_5.4.2_3.0_1724143900134.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kltn_coqe_vit5_total_spaol_v5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kltn_coqe_vit5_total_spaol_v5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kltn_coqe_vit5_total_spaol_v5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/ThuyNT03/KLTN_COQE_viT5_total_SPAOL_v5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_en.md b/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_en.md new file mode 100644 index 00000000000000..fcb72dddfdb331 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English kovr_t5 T5Transformer from Dm271 +author: John Snow Labs +name: kovr_t5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kovr_t5` is a English model originally trained by Dm271. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kovr_t5_en_5.4.2_3.0_1724117782436.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kovr_t5_en_5.4.2_3.0_1724117782436.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("kovr_t5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("kovr_t5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kovr_t5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|997.3 MB| + +## References + +https://huggingface.co/Dm271/Kovr_T5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_pipeline_en.md new file mode 100644 index 00000000000000..aaed867c585da4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-kovr_t5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English kovr_t5_pipeline pipeline T5Transformer from Dm271 +author: John Snow Labs +name: kovr_t5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`kovr_t5_pipeline` is a English model originally trained by Dm271. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/kovr_t5_pipeline_en_5.4.2_3.0_1724117830814.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/kovr_t5_pipeline_en_5.4.2_3.0_1724117830814.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("kovr_t5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("kovr_t5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|kovr_t5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|997.3 MB| + +## References + +https://huggingface.co/Dm271/Kovr_T5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_en.md b/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_en.md new file mode 100644 index 00000000000000..18d19cc6cfd687 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lat_long_1 T5Transformer from sahithya20 +author: John Snow Labs +name: lat_long_1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lat_long_1` is a English model originally trained by sahithya20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lat_long_1_en_5.4.2_3.0_1724134704321.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lat_long_1_en_5.4.2_3.0_1724134704321.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lat_long_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lat_long_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lat_long_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/sahithya20/lat_long_1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_pipeline_en.md new file mode 100644 index 00000000000000..041f0b40355a05 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-lat_long_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lat_long_1_pipeline pipeline T5Transformer from sahithya20 +author: John Snow Labs +name: lat_long_1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lat_long_1_pipeline` is a English model originally trained by sahithya20. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lat_long_1_pipeline_en_5.4.2_3.0_1724134723847.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lat_long_1_pipeline_en_5.4.2_3.0_1724134723847.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lat_long_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lat_long_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lat_long_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|332.2 MB| + +## References + +https://huggingface.co/sahithya20/lat_long_1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-legal_t5_small_trans_swedish_czech_small_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-legal_t5_small_trans_swedish_czech_small_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..b084b0d9c9669e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-legal_t5_small_trans_swedish_czech_small_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English legal_t5_small_trans_swedish_czech_small_finetuned_pipeline pipeline T5Transformer from SEBIS +author: John Snow Labs +name: legal_t5_small_trans_swedish_czech_small_finetuned_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`legal_t5_small_trans_swedish_czech_small_finetuned_pipeline` is a English model originally trained by SEBIS. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_swedish_czech_small_finetuned_pipeline_en_5.4.2_3.0_1724150889614.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/legal_t5_small_trans_swedish_czech_small_finetuned_pipeline_en_5.4.2_3.0_1724150889614.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("legal_t5_small_trans_swedish_czech_small_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("legal_t5_small_trans_swedish_czech_small_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|legal_t5_small_trans_swedish_czech_small_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.3 MB| + +## References + +https://huggingface.co/SEBIS/legal_t5_small_trans_sv_cs_small_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_en.md b/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_en.md new file mode 100644 index 00000000000000..61361305fc2a30 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English lite_llm_248 T5Transformer from currentlyexhausted +author: John Snow Labs +name: lite_llm_248 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lite_llm_248` is a English model originally trained by currentlyexhausted. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lite_llm_248_en_5.4.2_3.0_1724130387313.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lite_llm_248_en_5.4.2_3.0_1724130387313.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("lite_llm_248","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("lite_llm_248", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lite_llm_248| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/currentlyexhausted/lite-llm-248 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_pipeline_en.md new file mode 100644 index 00000000000000..1331b166e08771 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-lite_llm_248_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English lite_llm_248_pipeline pipeline T5Transformer from currentlyexhausted +author: John Snow Labs +name: lite_llm_248_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`lite_llm_248_pipeline` is a English model originally trained by currentlyexhausted. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/lite_llm_248_pipeline_en_5.4.2_3.0_1724130433914.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/lite_llm_248_pipeline_en_5.4.2_3.0_1724130433914.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("lite_llm_248_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("lite_llm_248_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|lite_llm_248_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/currentlyexhausted/lite-llm-248 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_en.md b/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_en.md new file mode 100644 index 00000000000000..3c3d7e908aa4a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English md_mt5_0109_v2 T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_0109_v2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_0109_v2` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_0109_v2_en_5.4.2_3.0_1724114927870.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_0109_v2_en_5.4.2_3.0_1724114927870.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("md_mt5_0109_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("md_mt5_0109_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_0109_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_0109_v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_pipeline_en.md new file mode 100644 index 00000000000000..1b5d0e0626656e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-md_mt5_0109_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English md_mt5_0109_v2_pipeline pipeline T5Transformer from Buseak +author: John Snow Labs +name: md_mt5_0109_v2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`md_mt5_0109_v2_pipeline` is a English model originally trained by Buseak. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/md_mt5_0109_v2_pipeline_en_5.4.2_3.0_1724115078633.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/md_mt5_0109_v2_pipeline_en_5.4.2_3.0_1724115078633.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("md_mt5_0109_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("md_mt5_0109_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|md_mt5_0109_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Buseak/md_mt5_0109_v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_en.md new file mode 100644 index 00000000000000..9b87bc6cf330fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_bangla_para_v1 T5Transformer from mHossain +author: John Snow Labs +name: mt5_base_bangla_para_v1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_bangla_para_v1` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_bangla_para_v1_en_5.4.2_3.0_1724142416239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_bangla_para_v1_en_5.4.2_3.0_1724142416239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_bangla_para_v1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_bangla_para_v1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_bangla_para_v1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/mHossain/mt5-base-bangla-para-v1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_pipeline_en.md new file mode 100644 index 00000000000000..631cbddc27bf8b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_bangla_para_v1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_bangla_para_v1_pipeline pipeline T5Transformer from mHossain +author: John Snow Labs +name: mt5_base_bangla_para_v1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_bangla_para_v1_pipeline` is a English model originally trained by mHossain. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_bangla_para_v1_pipeline_en_5.4.2_3.0_1724142592228.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_bangla_para_v1_pipeline_en_5.4.2_3.0_1724142592228.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_bangla_para_v1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_bangla_para_v1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_bangla_para_v1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.2 GB| + +## References + +https://huggingface.co/mHossain/mt5-base-bangla-para-v1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_en.md new file mode 100644 index 00000000000000..ac8dd75a46e63d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_cleaner T5Transformer from mojibaque +author: John Snow Labs +name: mt5_base_cleaner +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_cleaner` is a English model originally trained by mojibaque. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_cleaner_en_5.4.2_3.0_1724122395899.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_cleaner_en_5.4.2_3.0_1724122395899.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_cleaner","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_cleaner", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_cleaner| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/mojibaque/mt5-base-cleaner \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_pipeline_en.md new file mode 100644 index 00000000000000..88668c2bf06918 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_cleaner_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_cleaner_pipeline pipeline T5Transformer from mojibaque +author: John Snow Labs +name: mt5_base_cleaner_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_cleaner_pipeline` is a English model originally trained by mojibaque. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_cleaner_pipeline_en_5.4.2_3.0_1724122599899.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_cleaner_pipeline_en_5.4.2_3.0_1724122599899.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_cleaner_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_cleaner_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_cleaner_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/mojibaque/mt5-base-cleaner + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_en.md new file mode 100644 index 00000000000000..005e0c5e6857ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_coba_coba_coba T5Transformer from GhifSmile +author: John Snow Labs +name: mt5_base_coba_coba_coba +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_coba_coba_coba` is a English model originally trained by GhifSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_coba_coba_coba_en_5.4.2_3.0_1724137222400.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_coba_coba_coba_en_5.4.2_3.0_1724137222400.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_coba_coba_coba","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_coba_coba_coba", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_coba_coba_coba| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/GhifSmile/mt5-base-coba-coba-coba \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_pipeline_en.md new file mode 100644 index 00000000000000..50348bceb7e474 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_coba_coba_coba_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_coba_coba_coba_pipeline pipeline T5Transformer from GhifSmile +author: John Snow Labs +name: mt5_base_coba_coba_coba_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_coba_coba_coba_pipeline` is a English model originally trained by GhifSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_coba_coba_coba_pipeline_en_5.4.2_3.0_1724137436398.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_coba_coba_coba_pipeline_en_5.4.2_3.0_1724137436398.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_coba_coba_coba_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_coba_coba_coba_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_coba_coba_coba_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/GhifSmile/mt5-base-coba-coba-coba + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people_en.md new file mode 100644 index 00000000000000..8dd7b35d106fe4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people T5Transformer from himanshubeniwal +author: John Snow Labs +name: mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people` is a English model originally trained by himanshubeniwal. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people_en_5.4.2_3.0_1724146632645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people_en_5.4.2_3.0_1724146632645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_finetuned_kazakh_tonga_tonga_islands_english_dump_people| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.5 GB| + +## References + +https://huggingface.co/himanshubeniwal/mt5-base-finetuned-kk-to-en-dump-people \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_en.md new file mode 100644 index 00000000000000..cd0d6020c75bef --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_gramatika_final_e8_b16 T5Transformer from jeremyvictor +author: John Snow Labs +name: mt5_base_gramatika_final_e8_b16 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_gramatika_final_e8_b16` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_gramatika_final_e8_b16_en_5.4.2_3.0_1724117113919.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_gramatika_final_e8_b16_en_5.4.2_3.0_1724117113919.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_gramatika_final_e8_b16","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_gramatika_final_e8_b16", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_gramatika_final_e8_b16| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/jeremyvictor/mt5-base-gramatika-final-e8-b16 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_pipeline_en.md new file mode 100644 index 00000000000000..3c99291a8ee63d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_gramatika_final_e8_b16_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_gramatika_final_e8_b16_pipeline pipeline T5Transformer from jeremyvictor +author: John Snow Labs +name: mt5_base_gramatika_final_e8_b16_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_gramatika_final_e8_b16_pipeline` is a English model originally trained by jeremyvictor. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724117429158.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_gramatika_final_e8_b16_pipeline_en_5.4.2_3.0_1724117429158.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_gramatika_final_e8_b16_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_gramatika_final_e8_b16_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_gramatika_final_e8_b16_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/jeremyvictor/mt5-base-gramatika-final-e8-b16 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_en.md new file mode 100644 index 00000000000000..aeb287175acd88 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_jaquad_qag_trimmed_50000 T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_jaquad_qag_trimmed_50000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qag_trimmed_50000` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qag_trimmed_50000_en_5.4.2_3.0_1724150311079.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qag_trimmed_50000_en_5.4.2_3.0_1724150311079.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_jaquad_qag_trimmed_50000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_jaquad_qag_trimmed_50000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qag_trimmed_50000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-jaquad-qag-trimmed-50000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_pipeline_en.md new file mode 100644 index 00000000000000..9f0a4872790860 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qag_trimmed_50000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_jaquad_qag_trimmed_50000_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_jaquad_qag_trimmed_50000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qag_trimmed_50000_pipeline` is a English model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qag_trimmed_50000_pipeline_en_5.4.2_3.0_1724150376860.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qag_trimmed_50000_pipeline_en_5.4.2_3.0_1724150376860.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_jaquad_qag_trimmed_50000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_jaquad_qag_trimmed_50000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qag_trimmed_50000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-jaquad-qag-trimmed-50000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_en.md new file mode 100644 index 00000000000000..0958a8355929e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_jaquad_qg_trimmed_120000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_jaquad_qg_trimmed_120000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qg_trimmed_120000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_120000_en_5.4.2_3.0_1724126203610.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_120000_en_5.4.2_3.0_1724126203610.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_jaquad_qg_trimmed_120000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_jaquad_qg_trimmed_120000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qg_trimmed_120000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-jaquad-qg-trimmed-120000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_pipeline_en.md new file mode 100644 index 00000000000000..3d89b3f9ea7ebd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_jaquad_qg_trimmed_120000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_jaquad_qg_trimmed_120000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_jaquad_qg_trimmed_120000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_jaquad_qg_trimmed_120000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_120000_pipeline_en_5.4.2_3.0_1724126307169.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_jaquad_qg_trimmed_120000_pipeline_en_5.4.2_3.0_1724126307169.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_jaquad_qg_trimmed_120000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_jaquad_qg_trimmed_120000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_jaquad_qg_trimmed_120000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/research-backup/mt5-base-jaquad-qg-trimmed-120000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_cs.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_cs.md new file mode 100644 index 00000000000000..4daa7452e71bf5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_cs.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Czech mt5_base_multi_label_czech_iiib T5Transformer from chi2024 +author: John Snow Labs +name: mt5_base_multi_label_czech_iiib +date: 2024-08-20 +tags: [cs, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: cs +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_multi_label_czech_iiib` is a Czech model originally trained by chi2024. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_multi_label_czech_iiib_cs_5.4.2_3.0_1724125797747.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_multi_label_czech_iiib_cs_5.4.2_3.0_1724125797747.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_multi_label_czech_iiib","cs") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_multi_label_czech_iiib", "cs") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_multi_label_czech_iiib| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|cs| +|Size:|2.3 GB| + +## References + +https://huggingface.co/chi2024/mt5-base-multi-label-cs-iiib \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_pipeline_cs.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_pipeline_cs.md new file mode 100644 index 00000000000000..33b3dd6ca3312c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_multi_label_czech_iiib_pipeline_cs.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Czech mt5_base_multi_label_czech_iiib_pipeline pipeline T5Transformer from chi2024 +author: John Snow Labs +name: mt5_base_multi_label_czech_iiib_pipeline +date: 2024-08-20 +tags: [cs, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: cs +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_multi_label_czech_iiib_pipeline` is a Czech model originally trained by chi2024. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_multi_label_czech_iiib_pipeline_cs_5.4.2_3.0_1724126089775.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_multi_label_czech_iiib_pipeline_cs_5.4.2_3.0_1724126089775.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_multi_label_czech_iiib_pipeline", lang = "cs") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_multi_label_czech_iiib_pipeline", lang = "cs") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_multi_label_czech_iiib_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|cs| +|Size:|2.3 GB| + +## References + +https://huggingface.co/chi2024/mt5-base-multi-label-cs-iiib + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_en.md new file mode 100644 index 00000000000000..690e5c06acc6c1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_base_trimmed_italian_75000 T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_italian_75000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_italian_75000` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_75000_en_5.4.2_3.0_1724130535948.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_75000_en_5.4.2_3.0_1724130535948.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_trimmed_italian_75000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_trimmed_italian_75000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_italian_75000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|711.5 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-it-75000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_pipeline_en.md new file mode 100644 index 00000000000000..11cb6780d9b51f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_italian_75000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_trimmed_italian_75000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_italian_75000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_italian_75000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_75000_pipeline_en_5.4.2_3.0_1724130773106.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_italian_75000_pipeline_en_5.4.2_3.0_1724130773106.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_trimmed_italian_75000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_trimmed_italian_75000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_italian_75000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|711.5 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-it-75000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_japanese_60000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_japanese_60000_pipeline_en.md new file mode 100644 index 00000000000000..d0c9fe86d57577 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_trimmed_japanese_60000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_base_trimmed_japanese_60000_pipeline pipeline T5Transformer from research-backup +author: John Snow Labs +name: mt5_base_trimmed_japanese_60000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_trimmed_japanese_60000_pipeline` is a English model originally trained by research-backup. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_japanese_60000_pipeline_en_5.4.2_3.0_1724112116396.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_trimmed_japanese_60000_pipeline_en_5.4.2_3.0_1724112116396.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_trimmed_japanese_60000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_trimmed_japanese_60000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_trimmed_japanese_60000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|645.5 MB| + +## References + +https://huggingface.co/research-backup/mt5-base-trimmed-ja-60000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_pipeline_zh.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_pipeline_zh.md new file mode 100644 index 00000000000000..24d8c2afe8fbff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_pipeline_zh.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Chinese mt5_base_zhquad_qg_pipeline pipeline T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_zhquad_qg_pipeline +date: 2024-08-20 +tags: [zh, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_zhquad_qg_pipeline` is a Chinese model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_zhquad_qg_pipeline_zh_5.4.2_3.0_1724139306736.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_zhquad_qg_pipeline_zh_5.4.2_3.0_1724139306736.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_base_zhquad_qg_pipeline", lang = "zh") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_base_zhquad_qg_pipeline", lang = "zh") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_zhquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|zh| +|Size:|2.4 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-zhquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_zh.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_zh.md new file mode 100644 index 00000000000000..56295f120a7fc8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_base_zhquad_qg_zh.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Chinese mt5_base_zhquad_qg T5Transformer from lmqg +author: John Snow Labs +name: mt5_base_zhquad_qg +date: 2024-08-20 +tags: [zh, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: zh +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_base_zhquad_qg` is a Chinese model originally trained by lmqg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_base_zhquad_qg_zh_5.4.2_3.0_1724139120401.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_base_zhquad_qg_zh_5.4.2_3.0_1724139120401.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_base_zhquad_qg","zh") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_base_zhquad_qg", "zh") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_base_zhquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|zh| +|Size:|2.4 GB| + +## References + +https://huggingface.co/lmqg/mt5-base-zhquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_en.md new file mode 100644 index 00000000000000..0eaea79efc80a3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_baseline T5Transformer from kyle0518 +author: John Snow Labs +name: mt5_baseline +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_baseline` is a English model originally trained by kyle0518. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_baseline_en_5.4.2_3.0_1724146348054.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_baseline_en_5.4.2_3.0_1724146348054.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_baseline","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_baseline", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_baseline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/kyle0518/mt5_baseline \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_pipeline_en.md new file mode 100644 index 00000000000000..b59905340d6f87 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_baseline_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_baseline_pipeline pipeline T5Transformer from kyle0518 +author: John Snow Labs +name: mt5_baseline_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_baseline_pipeline` is a English model originally trained by kyle0518. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_baseline_pipeline_en_5.4.2_3.0_1724146510386.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_baseline_pipeline_en_5.4.2_3.0_1724146510386.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_baseline_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_baseline_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_baseline_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/kyle0518/mt5_baseline + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_en.md new file mode 100644 index 00000000000000..1ab3dbf4f53768 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_portuguese_rr_1000_v2 T5Transformer from MarianaLC +author: John Snow Labs +name: mt5_portuguese_rr_1000_v2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_portuguese_rr_1000_v2` is a English model originally trained by MarianaLC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_portuguese_rr_1000_v2_en_5.4.2_3.0_1724121456892.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_portuguese_rr_1000_v2_en_5.4.2_3.0_1724121456892.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_portuguese_rr_1000_v2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_portuguese_rr_1000_v2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_portuguese_rr_1000_v2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/MarianaLC/mt5-pt-rr-1000-v2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_pipeline_en.md new file mode 100644 index 00000000000000..e0a83e9a781667 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_portuguese_rr_1000_v2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_portuguese_rr_1000_v2_pipeline pipeline T5Transformer from MarianaLC +author: John Snow Labs +name: mt5_portuguese_rr_1000_v2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_portuguese_rr_1000_v2_pipeline` is a English model originally trained by MarianaLC. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_portuguese_rr_1000_v2_pipeline_en_5.4.2_3.0_1724121652177.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_portuguese_rr_1000_v2_pipeline_en_5.4.2_3.0_1724121652177.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_portuguese_rr_1000_v2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_portuguese_rr_1000_v2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_portuguese_rr_1000_v2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.4 GB| + +## References + +https://huggingface.co/MarianaLC/mt5-pt-rr-1000-v2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_pipeline_tr.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_pipeline_tr.md new file mode 100644 index 00000000000000..58f413d2057397 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_pipeline_tr.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Turkish mt5_small_3task_highlight_tquad2_pipeline pipeline T5Transformer from obss +author: John Snow Labs +name: mt5_small_3task_highlight_tquad2_pipeline +date: 2024-08-20 +tags: [tr, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: tr +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_3task_highlight_tquad2_pipeline` is a Turkish model originally trained by obss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_3task_highlight_tquad2_pipeline_tr_5.4.2_3.0_1724126530984.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_3task_highlight_tquad2_pipeline_tr_5.4.2_3.0_1724126530984.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_3task_highlight_tquad2_pipeline", lang = "tr") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_3task_highlight_tquad2_pipeline", lang = "tr") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_3task_highlight_tquad2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|tr| +|Size:|1.2 GB| + +## References + +https://huggingface.co/obss/mt5-small-3task-highlight-tquad2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_tr.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_tr.md new file mode 100644 index 00000000000000..900a34bf004b9d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_3task_highlight_tquad2_tr.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Turkish mt5_small_3task_highlight_tquad2 T5Transformer from obss +author: John Snow Labs +name: mt5_small_3task_highlight_tquad2 +date: 2024-08-20 +tags: [tr, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: tr +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_3task_highlight_tquad2` is a Turkish model originally trained by obss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_3task_highlight_tquad2_tr_5.4.2_3.0_1724126341414.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_3task_highlight_tquad2_tr_5.4.2_3.0_1724126341414.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_3task_highlight_tquad2","tr") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_3task_highlight_tquad2", "tr") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_3task_highlight_tquad2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|tr| +|Size:|1.2 GB| + +## References + +https://huggingface.co/obss/mt5-small-3task-highlight-tquad2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_en.md new file mode 100644 index 00000000000000..9444a881433e6f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_all_25000 T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_small_all_25000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_all_25000` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_all_25000_en_5.4.2_3.0_1724148013029.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_all_25000_en_5.4.2_3.0_1724148013029.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_all_25000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_all_25000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_all_25000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.4 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_small_all_25000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_pipeline_en.md new file mode 100644 index 00000000000000..5e217505d66387 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_all_25000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_all_25000_pipeline pipeline T5Transformer from santoshtyss +author: John Snow Labs +name: mt5_small_all_25000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_all_25000_pipeline` is a English model originally trained by santoshtyss. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_all_25000_pipeline_en_5.4.2_3.0_1724148127905.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_all_25000_pipeline_en_5.4.2_3.0_1724148127905.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_all_25000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_all_25000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_all_25000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.4 GB| + +## References + +https://huggingface.co/santoshtyss/mt5_small_all_25000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_en.md new file mode 100644 index 00000000000000..37acbbb7a4cb15 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_basque_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_basque_10k +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_basque_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_basque_10k_en_5.4.2_3.0_1724123661721.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_basque_10k_en_5.4.2_3.0_1724123661721.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_basque_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_basque_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_basque_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-eu-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_pipeline_en.md new file mode 100644 index 00000000000000..df4ee9519d553a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_basque_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_basque_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_basque_10k_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_basque_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_basque_10k_pipeline_en_5.4.2_3.0_1724123826120.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_basque_10k_pipeline_en_5.4.2_3.0_1724123826120.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_basque_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_basque_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_basque_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-eu-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_en.md new file mode 100644 index 00000000000000..3c592ded323179 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_czech_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_czech_10k +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_czech_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_czech_10k_en_5.4.2_3.0_1724145054624.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_czech_10k_en_5.4.2_3.0_1724145054624.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_czech_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_czech_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_czech_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-cs-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_pipeline_en.md new file mode 100644 index 00000000000000..fcef6b2fa5934d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_czech_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_czech_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_czech_10k_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_czech_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_czech_10k_pipeline_en_5.4.2_3.0_1724145208287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_czech_10k_pipeline_en_5.4.2_3.0_1724145208287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_czech_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_czech_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_czech_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-cs-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_en.md new file mode 100644 index 00000000000000..9881ac5d7eb997 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_english_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_english_10k +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_english_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_english_10k_en_5.4.2_3.0_1724135190578.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_english_10k_en_5.4.2_3.0_1724135190578.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_english_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_english_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_english_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-en-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_pipeline_en.md new file mode 100644 index 00000000000000..6366be9db74398 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_english_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_english_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_english_10k_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_english_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_english_10k_pipeline_en_5.4.2_3.0_1724135337355.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_english_10k_pipeline_en_5.4.2_3.0_1724135337355.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_english_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_english_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_english_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-en-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_en.md new file mode 100644 index 00000000000000..49fd932ac07496 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_chinese T5Transformer from yli418 +author: John Snow Labs +name: mt5_small_finetuned_amazon_chinese +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_chinese` is a English model originally trained by yli418. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_en_5.4.2_3.0_1724143783326.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_en_5.4.2_3.0_1724143783326.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_chinese","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_chinese", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_chinese| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/yli418/mt5-small-finetuned-amazon-zh \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_pipeline_en.md new file mode 100644 index 00000000000000..05059c4b2e1fe3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_chinese_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_chinese_pipeline pipeline T5Transformer from yli418 +author: John Snow Labs +name: mt5_small_finetuned_amazon_chinese_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_chinese_pipeline` is a English model originally trained by yli418. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_pipeline_en_5.4.2_3.0_1724143875315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_chinese_pipeline_en_5.4.2_3.0_1724143875315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_chinese_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_chinese_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_chinese_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/yli418/mt5-small-finetuned-amazon-zh + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_en.md new file mode 100644 index 00000000000000..0761442c238158 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_yuvscherbatov T5Transformer from yuvscherbatov +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_yuvscherbatov +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_yuvscherbatov` is a English model originally trained by yuvscherbatov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_yuvscherbatov_en_5.4.2_3.0_1724128632108.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_yuvscherbatov_en_5.4.2_3.0_1724128632108.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_yuvscherbatov","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_german_yuvscherbatov", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_yuvscherbatov| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/yuvscherbatov/mt5-small-finetuned-amazon-en-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline_en.md new file mode 100644 index 00000000000000..907ac1ad908af4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline pipeline T5Transformer from yuvscherbatov +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline` is a English model originally trained by yuvscherbatov. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline_en_5.4.2_3.0_1724128752540.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline_en_5.4.2_3.0_1724128752540.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_german_yuvscherbatov_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/yuvscherbatov/mt5-small-finetuned-amazon-en-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_en.md new file mode 100644 index 00000000000000..fa7ee2f6927482 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_accelerate_chisun T5Transformer from chisun +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_accelerate_chisun +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_accelerate_chisun` is a English model originally trained by chisun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_en_5.4.2_3.0_1724119816947.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_en_5.4.2_3.0_1724119816947.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_accelerate_chisun","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_accelerate_chisun", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_accelerate_chisun| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/chisun/mt5-small-finetuned-amazon-en-es-accelerate \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline_en.md new file mode 100644 index 00000000000000..c05f9bbdc6b7a6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline pipeline T5Transformer from chisun +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline` is a English model originally trained by chisun. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline_en_5.4.2_3.0_1724119966830.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline_en_5.4.2_3.0_1724119966830.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_accelerate_chisun_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/chisun/mt5-small-finetuned-amazon-en-es-accelerate + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_en.md new file mode 100644 index 00000000000000..7ce9fd115afba7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_cleverbrugger T5Transformer from cleverbrugger +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_cleverbrugger +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_cleverbrugger` is a English model originally trained by cleverbrugger. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cleverbrugger_en_5.4.2_3.0_1724117963975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cleverbrugger_en_5.4.2_3.0_1724117963975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_cleverbrugger","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_cleverbrugger", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_cleverbrugger| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/cleverbrugger/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline_en.md new file mode 100644 index 00000000000000..44a1c59afa08ca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline pipeline T5Transformer from cleverbrugger +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline` is a English model originally trained by cleverbrugger. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline_en_5.4.2_3.0_1724118065179.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline_en_5.4.2_3.0_1724118065179.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_cleverbrugger_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/cleverbrugger/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_en.md new file mode 100644 index 00000000000000..a7e27892ae00e0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_erfangc T5Transformer from erfangc +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_erfangc +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_erfangc` is a English model originally trained by erfangc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_erfangc_en_5.4.2_3.0_1724135598254.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_erfangc_en_5.4.2_3.0_1724135598254.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_erfangc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_erfangc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_erfangc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/erfangc/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline_en.md new file mode 100644 index 00000000000000..0ad158e254b847 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline pipeline T5Transformer from erfangc +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline` is a English model originally trained by erfangc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline_en_5.4.2_3.0_1724135689584.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline_en_5.4.2_3.0_1724135689584.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_erfangc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/erfangc/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_en.md new file mode 100644 index 00000000000000..f4b179727fd82b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_illuminarchie T5Transformer from Illuminarchie +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_illuminarchie +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_illuminarchie` is a English model originally trained by Illuminarchie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_illuminarchie_en_5.4.2_3.0_1724141070239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_illuminarchie_en_5.4.2_3.0_1724141070239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_illuminarchie","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_illuminarchie", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_illuminarchie| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Illuminarchie/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline_en.md new file mode 100644 index 00000000000000..2228e74ac78533 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline pipeline T5Transformer from Illuminarchie +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline` is a English model originally trained by Illuminarchie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline_en_5.4.2_3.0_1724141189354.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline_en_5.4.2_3.0_1724141189354.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_illuminarchie_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Illuminarchie/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_en.md new file mode 100644 index 00000000000000..3cdcea85f690d0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_tofunumber1 T5Transformer from TofuNumber1 +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_tofunumber1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_tofunumber1` is a English model originally trained by TofuNumber1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_tofunumber1_en_5.4.2_3.0_1724125733477.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_tofunumber1_en_5.4.2_3.0_1724125733477.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_tofunumber1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_tofunumber1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_tofunumber1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/TofuNumber1/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline_en.md new file mode 100644 index 00000000000000..36761c8cfc516b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline pipeline T5Transformer from TofuNumber1 +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline` is a English model originally trained by TofuNumber1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline_en_5.4.2_3.0_1724125833071.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline_en_5.4.2_3.0_1724125833071.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_tofunumber1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/TofuNumber1/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_en.md new file mode 100644 index 00000000000000..408110e82aa036 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_vicevk T5Transformer from ViceVk +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_vicevk +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_vicevk` is a English model originally trained by ViceVk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_vicevk_en_5.4.2_3.0_1724118144419.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_vicevk_en_5.4.2_3.0_1724118144419.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_vicevk","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_vicevk", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_vicevk| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ViceVk/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline_en.md new file mode 100644 index 00000000000000..4c4319af37120e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline pipeline T5Transformer from ViceVk +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline` is a English model originally trained by ViceVk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline_en_5.4.2_3.0_1724118251224.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline_en_5.4.2_3.0_1724118251224.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_vicevk_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/ViceVk/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_en.md new file mode 100644 index 00000000000000..ebcccfdd7f8c49 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_whatdhack T5Transformer from whatdhack +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_whatdhack +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_whatdhack` is a English model originally trained by whatdhack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_whatdhack_en_5.4.2_3.0_1724133985168.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_whatdhack_en_5.4.2_3.0_1724133985168.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_whatdhack","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_spanish_whatdhack", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_whatdhack| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/whatdhack/mt5-small-finetuned-amazon-en-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline_en.md new file mode 100644 index 00000000000000..48cc7b5786882c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline pipeline T5Transformer from whatdhack +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline` is a English model originally trained by whatdhack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline_en_5.4.2_3.0_1724134079150.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline_en_5.4.2_3.0_1724134079150.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_spanish_whatdhack_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/whatdhack/mt5-small-finetuned-amazon-en-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_en.md new file mode 100644 index 00000000000000..4212c6d46aa138 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_szymon T5Transformer from Szymon +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_szymon +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_szymon` is a English model originally trained by Szymon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_szymon_en_5.4.2_3.0_1724139274287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_szymon_en_5.4.2_3.0_1724139274287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_szymon","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_amazon_english_szymon", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_szymon| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Szymon/mt5-small-finetuned-amazon-en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_pipeline_en.md new file mode 100644 index 00000000000000..f0f372aa14a99c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_amazon_english_szymon_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_amazon_english_szymon_pipeline pipeline T5Transformer from Szymon +author: John Snow Labs +name: mt5_small_finetuned_amazon_english_szymon_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_amazon_english_szymon_pipeline` is a English model originally trained by Szymon. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_szymon_pipeline_en_5.4.2_3.0_1724139368936.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_amazon_english_szymon_pipeline_en_5.4.2_3.0_1724139368936.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_amazon_english_szymon_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_amazon_english_szymon_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_amazon_english_szymon_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Szymon/mt5-small-finetuned-amazon-en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_en.md new file mode 100644 index 00000000000000..e8554be76845f2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_cnn_news T5Transformer from shivaniNK8 +author: John Snow Labs +name: mt5_small_finetuned_cnn_news +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_cnn_news` is a English model originally trained by shivaniNK8. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_cnn_news_en_5.4.2_3.0_1724117128020.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_cnn_news_en_5.4.2_3.0_1724117128020.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_cnn_news","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_cnn_news", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_cnn_news| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/shivaniNK8/mt5-small-finetuned-cnn-news \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_pipeline_en.md new file mode 100644 index 00000000000000..d1e88219ba6d81 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_cnn_news_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_cnn_news_pipeline pipeline T5Transformer from shivaniNK8 +author: John Snow Labs +name: mt5_small_finetuned_cnn_news_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_cnn_news_pipeline` is a English model originally trained by shivaniNK8. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_cnn_news_pipeline_en_5.4.2_3.0_1724117250369.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_cnn_news_pipeline_en_5.4.2_3.0_1724117250369.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_cnn_news_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_cnn_news_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_cnn_news_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/shivaniNK8/mt5-small-finetuned-cnn-news + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_en.md new file mode 100644 index 00000000000000..df05006934286a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_english_tonga_tonga_islands_korean_101 T5Transformer from alphahg +author: John Snow Labs +name: mt5_small_finetuned_english_tonga_tonga_islands_korean_101 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_english_tonga_tonga_islands_korean_101` is a English model originally trained by alphahg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_korean_101_en_5.4.2_3.0_1724131316982.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_korean_101_en_5.4.2_3.0_1724131316982.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_english_tonga_tonga_islands_korean_101","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_english_tonga_tonga_islands_korean_101", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_english_tonga_tonga_islands_korean_101| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/alphahg/mt5-small-finetuned-en-to-ko-101 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline_en.md new file mode 100644 index 00000000000000..c54ba63b702a53 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline pipeline T5Transformer from alphahg +author: John Snow Labs +name: mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline` is a English model originally trained by alphahg. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline_en_5.4.2_3.0_1724131586689.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline_en_5.4.2_3.0_1724131586689.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_english_tonga_tonga_islands_korean_101_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/alphahg/mt5-small-finetuned-en-to-ko-101 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_en.md new file mode 100644 index 00000000000000..953f301fbe52f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_finetuned_multihead_230209_test3 T5Transformer from HealthTeam +author: John Snow Labs +name: mt5_small_finetuned_multihead_230209_test3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_multihead_230209_test3` is a English model originally trained by HealthTeam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multihead_230209_test3_en_5.4.2_3.0_1724129098131.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multihead_230209_test3_en_5.4.2_3.0_1724129098131.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_finetuned_multihead_230209_test3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_finetuned_multihead_230209_test3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_multihead_230209_test3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/HealthTeam/mt5-small-finetuned-MultiHead-230209-test3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_pipeline_en.md new file mode 100644 index 00000000000000..4c517d6d9af2c0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_finetuned_multihead_230209_test3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_finetuned_multihead_230209_test3_pipeline pipeline T5Transformer from HealthTeam +author: John Snow Labs +name: mt5_small_finetuned_multihead_230209_test3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_finetuned_multihead_230209_test3_pipeline` is a English model originally trained by HealthTeam. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multihead_230209_test3_pipeline_en_5.4.2_3.0_1724129223835.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_finetuned_multihead_230209_test3_pipeline_en_5.4.2_3.0_1724129223835.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_finetuned_multihead_230209_test3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_finetuned_multihead_230209_test3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_finetuned_multihead_230209_test3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/HealthTeam/mt5-small-finetuned-MultiHead-230209-test3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_en.md new file mode 100644 index 00000000000000..69cabae8e28c2a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_ibn_shaddad_v4 T5Transformer from Ahmed007 +author: John Snow Labs +name: mt5_small_ibn_shaddad_v4 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ibn_shaddad_v4` is a English model originally trained by Ahmed007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ibn_shaddad_v4_en_5.4.2_3.0_1724125086168.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ibn_shaddad_v4_en_5.4.2_3.0_1724125086168.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_ibn_shaddad_v4","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_ibn_shaddad_v4", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ibn_shaddad_v4| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Ahmed007/mt5-small-ibn-Shaddad-v4 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_pipeline_en.md new file mode 100644 index 00000000000000..13dbcd2fb8f219 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_ibn_shaddad_v4_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_ibn_shaddad_v4_pipeline pipeline T5Transformer from Ahmed007 +author: John Snow Labs +name: mt5_small_ibn_shaddad_v4_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_ibn_shaddad_v4_pipeline` is a English model originally trained by Ahmed007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_ibn_shaddad_v4_pipeline_en_5.4.2_3.0_1724125181308.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_ibn_shaddad_v4_pipeline_en_5.4.2_3.0_1724125181308.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_ibn_shaddad_v4_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_ibn_shaddad_v4_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_ibn_shaddad_v4_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/Ahmed007/mt5-small-ibn-Shaddad-v4 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_en.md new file mode 100644 index 00000000000000..167605c11bf3ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_malayalam_10k T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_malayalam_10k +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_malayalam_10k` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_malayalam_10k_en_5.4.2_3.0_1724126944350.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_malayalam_10k_en_5.4.2_3.0_1724126944350.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_malayalam_10k","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_malayalam_10k", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_malayalam_10k| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ml-10k \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_pipeline_en.md new file mode 100644 index 00000000000000..79f827b23b9c2e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_malayalam_10k_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_malayalam_10k_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: mt5_small_malayalam_10k_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_malayalam_10k_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_malayalam_10k_pipeline_en_5.4.2_3.0_1724127102520.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_malayalam_10k_pipeline_en_5.4.2_3.0_1724127102520.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_malayalam_10k_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_malayalam_10k_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_malayalam_10k_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/KaiNylund/mt5-small-ml-10k + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_en.md new file mode 100644 index 00000000000000..d536723b73bbd3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_nc16_400_deen T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_400_deen +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_400_deen` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_400_deen_en_5.4.2_3.0_1724130975867.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_400_deen_en_5.4.2_3.0_1724130975867.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_nc16_400_deen","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_nc16_400_deen", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_400_deen| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-400-deen \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_pipeline_en.md new file mode 100644 index 00000000000000..1b0b88c57a97ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_nc16_400_deen_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_nc16_400_deen_pipeline pipeline T5Transformer from leukas +author: John Snow Labs +name: mt5_small_nc16_400_deen_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_nc16_400_deen_pipeline` is a English model originally trained by leukas. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_400_deen_pipeline_en_5.4.2_3.0_1724131166654.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_nc16_400_deen_pipeline_en_5.4.2_3.0_1724131166654.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_nc16_400_deen_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_nc16_400_deen_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_nc16_400_deen_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/leukas/mt5-small-nc16-400-deen + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_en.md new file mode 100644 index 00000000000000..e7aa21a5c00edb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_news_final T5Transformer from ThatGuyVanquish +author: John Snow Labs +name: mt5_small_news_final +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_news_final` is a English model originally trained by ThatGuyVanquish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_news_final_en_5.4.2_3.0_1724123636574.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_news_final_en_5.4.2_3.0_1724123636574.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_news_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_news_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_news_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/ThatGuyVanquish/mt5-small-news-final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_pipeline_en.md new file mode 100644 index 00000000000000..5a480945b5f3cf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_news_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_news_final_pipeline pipeline T5Transformer from ThatGuyVanquish +author: John Snow Labs +name: mt5_small_news_final_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_news_final_pipeline` is a English model originally trained by ThatGuyVanquish. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_news_final_pipeline_en_5.4.2_3.0_1724123925826.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_news_final_pipeline_en_5.4.2_3.0_1724123925826.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_news_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_news_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_news_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|819.8 MB| + +## References + +https://huggingface.co/ThatGuyVanquish/mt5-small-news-final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_en.md new file mode 100644 index 00000000000000..9f968b97d3a0c4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_sport T5Transformer from toanduc +author: John Snow Labs +name: mt5_small_sport +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_sport` is a English model originally trained by toanduc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_sport_en_5.4.2_3.0_1724135125946.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_sport_en_5.4.2_3.0_1724135125946.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_sport","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_sport", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_sport| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/toanduc/mt5-small-sport \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_pipeline_en.md new file mode 100644 index 00000000000000..9e7aad0313c9b0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_sport_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_sport_pipeline pipeline T5Transformer from toanduc +author: John Snow Labs +name: mt5_small_sport_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_sport_pipeline` is a English model originally trained by toanduc. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_sport_pipeline_en_5.4.2_3.0_1724135278985.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_sport_pipeline_en_5.4.2_3.0_1724135278985.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_sport_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_sport_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_sport_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.2 GB| + +## References + +https://huggingface.co/toanduc/mt5-small-sport + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_en.md new file mode 100644 index 00000000000000..9dc4196592813c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_squad_qa_trimmed_english_10000 T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_squad_qa_trimmed_english_10000 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qa_trimmed_english_10000` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_10000_en_5.4.2_3.0_1724139087925.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_10000_en_5.4.2_3.0_1724139087925.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_squad_qa_trimmed_english_10000","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_squad_qa_trimmed_english_10000", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qa_trimmed_english_10000| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|224.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-squad-qa-trimmed-en-10000 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_pipeline_en.md new file mode 100644 index 00000000000000..5d1ec2ef9c3e13 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_squad_qa_trimmed_english_10000_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_squad_qa_trimmed_english_10000_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_squad_qa_trimmed_english_10000_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_squad_qa_trimmed_english_10000_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_10000_pipeline_en_5.4.2_3.0_1724139098679.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_squad_qa_trimmed_english_10000_pipeline_en_5.4.2_3.0_1724139098679.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_squad_qa_trimmed_english_10000_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_squad_qa_trimmed_english_10000_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_squad_qa_trimmed_english_10000_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|224.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-squad-qa-trimmed-en-10000 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_it.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_it.md new file mode 100644 index 00000000000000..3dcdcf74759482 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_it.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_10000_itquad_qg T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_10000_itquad_qg +date: 2024-08-20 +tags: [it, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_10000_itquad_qg` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_10000_itquad_qg_it_5.4.2_3.0_1724146436127.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_10000_itquad_qg_it_5.4.2_3.0_1724146436127.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_10000_itquad_qg","it") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian_10000_itquad_qg", "it") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_10000_itquad_qg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|it| +|Size:|224.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-10000-itquad-qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_pipeline_it.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_pipeline_it.md new file mode 100644 index 00000000000000..118b794732cf24 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_10000_itquad_qg_pipeline_it.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Italian mt5_small_trimmed_italian_10000_itquad_qg_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_10000_itquad_qg_pipeline +date: 2024-08-20 +tags: [it, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: it +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_10000_itquad_qg_pipeline` is a Italian model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_10000_itquad_qg_pipeline_it_5.4.2_3.0_1724146447245.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_10000_itquad_qg_pipeline_it_5.4.2_3.0_1724146447245.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_10000_itquad_qg_pipeline", lang = "it") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_10000_itquad_qg_pipeline", lang = "it") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_10000_itquad_qg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|it| +|Size:|224.5 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it-10000-itquad-qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_en.md new file mode 100644 index 00000000000000..b662aaa2b2fa49 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English mt5_small_trimmed_italian T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_en_5.4.2_3.0_1724131263672.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_en_5.4.2_3.0_1724131263672.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_italian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_italian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|412.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_pipeline_en.md new file mode 100644 index 00000000000000..2e7e35c8d86cd1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_italian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English mt5_small_trimmed_italian_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_italian_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_italian_pipeline` is a English model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_pipeline_en_5.4.2_3.0_1724131398425.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_italian_pipeline_en_5.4.2_3.0_1724131398425.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_italian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_italian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_italian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|412.4 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-it + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_ko.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_ko.md new file mode 100644 index 00000000000000..c9de1807a3ec8b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_ko.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Korean mt5_small_trimmed_korean_10000_koquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_10000_koquad_qa +date: 2024-08-20 +tags: [ko, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_10000_koquad_qa` is a Korean model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_10000_koquad_qa_ko_5.4.2_3.0_1724136359049.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_10000_koquad_qa_ko_5.4.2_3.0_1724136359049.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_10000_koquad_qa","ko") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_10000_koquad_qa", "ko") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_10000_koquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ko| +|Size:|221.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-10000-koquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_pipeline_ko.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_pipeline_ko.md new file mode 100644 index 00000000000000..8aefa3b0503cf7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_10000_koquad_qa_pipeline_ko.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Korean mt5_small_trimmed_korean_10000_koquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_10000_koquad_qa_pipeline +date: 2024-08-20 +tags: [ko, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_10000_koquad_qa_pipeline` is a Korean model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_10000_koquad_qa_pipeline_ko_5.4.2_3.0_1724136369950.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_10000_koquad_qa_pipeline_ko_5.4.2_3.0_1724136369950.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_korean_10000_koquad_qa_pipeline", lang = "ko") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_korean_10000_koquad_qa_pipeline", lang = "ko") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_10000_koquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ko| +|Size:|221.2 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-10000-koquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_ko.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_ko.md new file mode 100644 index 00000000000000..ba21dd022ccc59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_ko.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Korean mt5_small_trimmed_korean_60000_koquad_qa T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_60000_koquad_qa +date: 2024-08-20 +tags: [ko, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_60000_koquad_qa` is a Korean model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_60000_koquad_qa_ko_5.4.2_3.0_1724112176015.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_60000_koquad_qa_ko_5.4.2_3.0_1724112176015.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_60000_koquad_qa","ko") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("mt5_small_trimmed_korean_60000_koquad_qa", "ko") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_60000_koquad_qa| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ko| +|Size:|448.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-60000-koquad-qa \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_pipeline_ko.md b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_pipeline_ko.md new file mode 100644 index 00000000000000..0d5e9cf760758b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-mt5_small_trimmed_korean_60000_koquad_qa_pipeline_ko.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Korean mt5_small_trimmed_korean_60000_koquad_qa_pipeline pipeline T5Transformer from vocabtrimmer +author: John Snow Labs +name: mt5_small_trimmed_korean_60000_koquad_qa_pipeline +date: 2024-08-20 +tags: [ko, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ko +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`mt5_small_trimmed_korean_60000_koquad_qa_pipeline` is a Korean model originally trained by vocabtrimmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_60000_koquad_qa_pipeline_ko_5.4.2_3.0_1724112202606.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/mt5_small_trimmed_korean_60000_koquad_qa_pipeline_ko_5.4.2_3.0_1724112202606.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("mt5_small_trimmed_korean_60000_koquad_qa_pipeline", lang = "ko") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("mt5_small_trimmed_korean_60000_koquad_qa_pipeline", lang = "ko") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|mt5_small_trimmed_korean_60000_koquad_qa_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ko| +|Size:|448.7 MB| + +## References + +https://huggingface.co/vocabtrimmer/mt5-small-trimmed-ko-60000-koquad-qa + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-net5_base_ne.md b/docs/_posts/ahmedlone127/2024-08-20-net5_base_ne.md new file mode 100644 index 00000000000000..50e7deb1b51025 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-net5_base_ne.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Nepali (macrolanguage) net5_base T5Transformer from Angeldahal404 +author: John Snow Labs +name: net5_base +date: 2024-08-20 +tags: [ne, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ne +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`net5_base` is a Nepali (macrolanguage) model originally trained by Angeldahal404. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/net5_base_ne_5.4.2_3.0_1724112850977.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/net5_base_ne_5.4.2_3.0_1724112850977.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("net5_base","ne") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("net5_base", "ne") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|net5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ne| +|Size:|462.5 MB| + +## References + +https://huggingface.co/Angeldahal404/net5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-net5_base_pipeline_ne.md b/docs/_posts/ahmedlone127/2024-08-20-net5_base_pipeline_ne.md new file mode 100644 index 00000000000000..0a1f82fb107c86 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-net5_base_pipeline_ne.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Nepali (macrolanguage) net5_base_pipeline pipeline T5Transformer from Angeldahal404 +author: John Snow Labs +name: net5_base_pipeline +date: 2024-08-20 +tags: [ne, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ne +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`net5_base_pipeline` is a Nepali (macrolanguage) model originally trained by Angeldahal404. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/net5_base_pipeline_ne_5.4.2_3.0_1724113012083.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/net5_base_pipeline_ne_5.4.2_3.0_1724113012083.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("net5_base_pipeline", lang = "ne") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("net5_base_pipeline", lang = "ne") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|net5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ne| +|Size:|462.5 MB| + +## References + +https://huggingface.co/Angeldahal404/net5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_en.md b/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_en.md new file mode 100644 index 00000000000000..1eac0e6f2739dd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English nlp_a3_2c_trained T5Transformer from samridhgirdhar +author: John Snow Labs +name: nlp_a3_2c_trained +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nlp_a3_2c_trained` is a English model originally trained by samridhgirdhar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nlp_a3_2c_trained_en_5.4.2_3.0_1724130936757.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nlp_a3_2c_trained_en_5.4.2_3.0_1724130936757.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("nlp_a3_2c_trained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("nlp_a3_2c_trained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nlp_a3_2c_trained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/samridhgirdhar/NLP_A3_2C_TRAINED \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_pipeline_en.md new file mode 100644 index 00000000000000..a3572cffbe3e95 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-nlp_a3_2c_trained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English nlp_a3_2c_trained_pipeline pipeline T5Transformer from samridhgirdhar +author: John Snow Labs +name: nlp_a3_2c_trained_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`nlp_a3_2c_trained_pipeline` is a English model originally trained by samridhgirdhar. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/nlp_a3_2c_trained_pipeline_en_5.4.2_3.0_1724130953253.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/nlp_a3_2c_trained_pipeline_en_5.4.2_3.0_1724130953253.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("nlp_a3_2c_trained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("nlp_a3_2c_trained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|nlp_a3_2c_trained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/samridhgirdhar/NLP_A3_2C_TRAINED + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_en.md b/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_en.md new file mode 100644 index 00000000000000..78d0ef046f4db0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English norgec_mt5 T5Transformer from MatiasJ +author: John Snow Labs +name: norgec_mt5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`norgec_mt5` is a English model originally trained by MatiasJ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/norgec_mt5_en_5.4.2_3.0_1724117939339.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/norgec_mt5_en_5.4.2_3.0_1724117939339.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("norgec_mt5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("norgec_mt5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|norgec_mt5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/MatiasJ/norgec_mt5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_pipeline_en.md new file mode 100644 index 00000000000000..da14b87331c64c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-norgec_mt5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English norgec_mt5_pipeline pipeline T5Transformer from MatiasJ +author: John Snow Labs +name: norgec_mt5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`norgec_mt5_pipeline` is a English model originally trained by MatiasJ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/norgec_mt5_pipeline_en_5.4.2_3.0_1724118101315.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/norgec_mt5_pipeline_en_5.4.2_3.0_1724118101315.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("norgec_mt5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("norgec_mt5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|norgec_mt5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.3 GB| + +## References + +https://huggingface.co/MatiasJ/norgec_mt5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_en.md b/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_en.md new file mode 100644 index 00000000000000..8040e02e183872 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English norwegian_sum_t5_6 T5Transformer from alraisi +author: John Snow Labs +name: norwegian_sum_t5_6 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`norwegian_sum_t5_6` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/norwegian_sum_t5_6_en_5.4.2_3.0_1724146549821.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/norwegian_sum_t5_6_en_5.4.2_3.0_1724146549821.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("norwegian_sum_t5_6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("norwegian_sum_t5_6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|norwegian_sum_t5_6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.3 MB| + +## References + +https://huggingface.co/alraisi/no-sum-t5-6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_pipeline_en.md new file mode 100644 index 00000000000000..672722578b360e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-norwegian_sum_t5_6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English norwegian_sum_t5_6_pipeline pipeline T5Transformer from alraisi +author: John Snow Labs +name: norwegian_sum_t5_6_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`norwegian_sum_t5_6_pipeline` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/norwegian_sum_t5_6_pipeline_en_5.4.2_3.0_1724146568620.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/norwegian_sum_t5_6_pipeline_en_5.4.2_3.0_1724146568620.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("norwegian_sum_t5_6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("norwegian_sum_t5_6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|norwegian_sum_t5_6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.3 MB| + +## References + +https://huggingface.co/alraisi/no-sum-t5-6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-panl_sql_en.md b/docs/_posts/ahmedlone127/2024-08-20-panl_sql_en.md new file mode 100644 index 00000000000000..c3a0a01950e1b2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-panl_sql_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English panl_sql T5Transformer from obada-jaras +author: John Snow Labs +name: panl_sql +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`panl_sql` is a English model originally trained by obada-jaras. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/panl_sql_en_5.4.2_3.0_1724129284318.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/panl_sql_en_5.4.2_3.0_1724129284318.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("panl_sql","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("panl_sql", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|panl_sql| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/obada-jaras/PANL_SQL \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-panl_sql_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-panl_sql_pipeline_en.md new file mode 100644 index 00000000000000..8cb202fec5a8c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-panl_sql_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English panl_sql_pipeline pipeline T5Transformer from obada-jaras +author: John Snow Labs +name: panl_sql_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`panl_sql_pipeline` is a English model originally trained by obada-jaras. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/panl_sql_pipeline_en_5.4.2_3.0_1724129366945.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/panl_sql_pipeline_en_5.4.2_3.0_1724129366945.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("panl_sql_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("panl_sql_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|panl_sql_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.7 GB| + +## References + +https://huggingface.co/obada-jaras/PANL_SQL + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_en.md b/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_en.md new file mode 100644 index 00000000000000..58dfd7cc497eaa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English paraphrase_generation_final T5Transformer from sharifMunna +author: John Snow Labs +name: paraphrase_generation_final +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`paraphrase_generation_final` is a English model originally trained by sharifMunna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/paraphrase_generation_final_en_5.4.2_3.0_1724114711657.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/paraphrase_generation_final_en_5.4.2_3.0_1724114711657.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("paraphrase_generation_final","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("paraphrase_generation_final", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|paraphrase_generation_final| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sharifMunna/paraphrase_generation_final \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_pipeline_en.md new file mode 100644 index 00000000000000..cc4b03af015031 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-paraphrase_generation_final_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English paraphrase_generation_final_pipeline pipeline T5Transformer from sharifMunna +author: John Snow Labs +name: paraphrase_generation_final_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`paraphrase_generation_final_pipeline` is a English model originally trained by sharifMunna. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/paraphrase_generation_final_pipeline_en_5.4.2_3.0_1724114774784.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/paraphrase_generation_final_pipeline_en_5.4.2_3.0_1724114774784.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("paraphrase_generation_final_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("paraphrase_generation_final_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|paraphrase_generation_final_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/sharifMunna/paraphrase_generation_final + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_en.md b/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_en.md new file mode 100644 index 00000000000000..3a622cb3c951ad --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English practical_nlp_course_6 T5Transformer from trtd56 +author: John Snow Labs +name: practical_nlp_course_6 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`practical_nlp_course_6` is a English model originally trained by trtd56. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/practical_nlp_course_6_en_5.4.2_3.0_1724129725213.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/practical_nlp_course_6_en_5.4.2_3.0_1724129725213.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("practical_nlp_course_6","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("practical_nlp_course_6", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|practical_nlp_course_6| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|314.3 MB| + +## References + +https://huggingface.co/trtd56/practical_nlp_course_6 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_pipeline_en.md new file mode 100644 index 00000000000000..4347d2288e8bf3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-practical_nlp_course_6_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English practical_nlp_course_6_pipeline pipeline T5Transformer from trtd56 +author: John Snow Labs +name: practical_nlp_course_6_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`practical_nlp_course_6_pipeline` is a English model originally trained by trtd56. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/practical_nlp_course_6_pipeline_en_5.4.2_3.0_1724129747963.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/practical_nlp_course_6_pipeline_en_5.4.2_3.0_1724129747963.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("practical_nlp_course_6_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("practical_nlp_course_6_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|practical_nlp_course_6_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|314.3 MB| + +## References + +https://huggingface.co/trtd56/practical_nlp_course_6 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_en.md b/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_en.md new file mode 100644 index 00000000000000..4c41e513276226 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English pretrain_law_model_vit5_version3 T5Transformer from KingLTD +author: John Snow Labs +name: pretrain_law_model_vit5_version3 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pretrain_law_model_vit5_version3` is a English model originally trained by KingLTD. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pretrain_law_model_vit5_version3_en_5.4.2_3.0_1724144739612.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pretrain_law_model_vit5_version3_en_5.4.2_3.0_1724144739612.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("pretrain_law_model_vit5_version3","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("pretrain_law_model_vit5_version3", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pretrain_law_model_vit5_version3| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/KingLTD/pretrain_Law_model_vit5_version3 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_pipeline_en.md new file mode 100644 index 00000000000000..5c4d0ec53eb6ea --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-pretrain_law_model_vit5_version3_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English pretrain_law_model_vit5_version3_pipeline pipeline T5Transformer from KingLTD +author: John Snow Labs +name: pretrain_law_model_vit5_version3_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`pretrain_law_model_vit5_version3_pipeline` is a English model originally trained by KingLTD. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/pretrain_law_model_vit5_version3_pipeline_en_5.4.2_3.0_1724144797444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/pretrain_law_model_vit5_version3_pipeline_en_5.4.2_3.0_1724144797444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("pretrain_law_model_vit5_version3_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("pretrain_law_model_vit5_version3_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|pretrain_law_model_vit5_version3_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/KingLTD/pretrain_Law_model_vit5_version3 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_en.md b/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_en.md new file mode 100644 index 00000000000000..182e25555ea1f7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qg_supercat666 T5Transformer from supercat666 +author: John Snow Labs +name: qg_supercat666 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qg_supercat666` is a English model originally trained by supercat666. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qg_supercat666_en_5.4.2_3.0_1724134760957.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qg_supercat666_en_5.4.2_3.0_1724134760957.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qg_supercat666","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qg_supercat666", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qg_supercat666| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.5 GB| + +## References + +https://huggingface.co/supercat666/qg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_pipeline_en.md new file mode 100644 index 00000000000000..4ce670c98c9895 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-qg_supercat666_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qg_supercat666_pipeline pipeline T5Transformer from supercat666 +author: John Snow Labs +name: qg_supercat666_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qg_supercat666_pipeline` is a English model originally trained by supercat666. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qg_supercat666_pipeline_en_5.4.2_3.0_1724134920509.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qg_supercat666_pipeline_en_5.4.2_3.0_1724134920509.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qg_supercat666_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qg_supercat666_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qg_supercat666_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.5 GB| + +## References + +https://huggingface.co/supercat666/qg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_en.md b/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_en.md new file mode 100644 index 00000000000000..66974ebbb65a1a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English qm_sum_flan_t5_large T5Transformer from iamanavk +author: John Snow Labs +name: qm_sum_flan_t5_large +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qm_sum_flan_t5_large` is a English model originally trained by iamanavk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qm_sum_flan_t5_large_en_5.4.2_3.0_1724112801763.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qm_sum_flan_t5_large_en_5.4.2_3.0_1724112801763.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("qm_sum_flan_t5_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("qm_sum_flan_t5_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qm_sum_flan_t5_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/iamanavk/qm_sum_flan_t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_pipeline_en.md new file mode 100644 index 00000000000000..0d1111f0ed362a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-qm_sum_flan_t5_large_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English qm_sum_flan_t5_large_pipeline pipeline T5Transformer from iamanavk +author: John Snow Labs +name: qm_sum_flan_t5_large_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`qm_sum_flan_t5_large_pipeline` is a English model originally trained by iamanavk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/qm_sum_flan_t5_large_pipeline_en_5.4.2_3.0_1724112952656.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/qm_sum_flan_t5_large_pipeline_en_5.4.2_3.0_1724112952656.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("qm_sum_flan_t5_large_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("qm_sum_flan_t5_large_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|qm_sum_flan_t5_large_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/iamanavk/qm_sum_flan_t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_en.md b/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_en.md new file mode 100644 index 00000000000000..9e6f3b4a737700 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_en.md @@ -0,0 +1,92 @@ +--- +layout: model +title: English question_answer_model DistilBertForQuestionAnswering from hameersiddique +author: John Snow Labs +name: question_answer_model +date: 2024-08-20 +tags: [distilbert, en, open_source, question_answering, onnx] +task: Question Answering +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained DistilBertForQuestionAnswering model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_answer_model` is a English model originally trained by hameersiddique. + +## Predicted Entities + + + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_answer_model_en_5.4.2_3.0_1724142857578.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_answer_model_en_5.4.2_3.0_1724142857578.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python +document_assembler = MultiDocumentAssembler() \ + .setInputCol(["question", "context"]) \ + .setOutputCol(["document_question", "document_context"]) + + +spanClassifier = DistilBertForQuestionAnswering.pretrained("question_answer_model","en") \ + .setInputCols(["document_question","document_context"]) \ + .setOutputCol("answer") + +pipeline = Pipeline().setStages([document_assembler, spanClassifier]) + +pipelineModel = pipeline.fit(data) + +pipelineDF = pipelineModel.transform(data) +``` +```scala +val document_assembler = new MultiDocumentAssembler() + .setInputCol(Array("question", "context")) + .setOutputCol(Array("document_question", "document_context")) + +val spanClassifier = DistilBertForQuestionAnswering + .pretrained("question_answer_model", "en") + .setInputCols(Array("document_question","document_context")) + .setOutputCol("answer") + +val pipeline = new Pipeline().setStages(Array(document_assembler, spanClassifier)) + +val pipelineModel = pipeline.fit(data) + +val pipelineDF = pipelineModel.transform(data) +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_answer_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|319.9 MB| + +## References + +References + +https://huggingface.co/hameersiddique/question_answer_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_pipeline_en.md new file mode 100644 index 00000000000000..9e3adef90f10a3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-question_answer_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English question_answer_model_pipeline pipeline T5Transformer from liorfieldwire +author: John Snow Labs +name: question_answer_model_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_answer_model_pipeline` is a English model originally trained by liorfieldwire. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_answer_model_pipeline_en_5.4.2_3.0_1724142879474.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_answer_model_pipeline_en_5.4.2_3.0_1724142879474.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("question_answer_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("question_answer_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_answer_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|319.9 MB| + +## References + +https://huggingface.co/liorfieldwire/question_answer_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_en.md b/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_en.md new file mode 100644 index 00000000000000..378c26a74749e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English question_model_t5_tokenizer T5Transformer from Mayank393 +author: John Snow Labs +name: question_model_t5_tokenizer +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_model_t5_tokenizer` is a English model originally trained by Mayank393. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_model_t5_tokenizer_en_5.4.2_3.0_1724135486541.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_model_t5_tokenizer_en_5.4.2_3.0_1724135486541.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("question_model_t5_tokenizer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("question_model_t5_tokenizer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_model_t5_tokenizer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.3 MB| + +## References + +https://huggingface.co/Mayank393/Question_Model_T5_Tokenizer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_pipeline_en.md new file mode 100644 index 00000000000000..471463ac5ca4e7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-question_model_t5_tokenizer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English question_model_t5_tokenizer_pipeline pipeline T5Transformer from Mayank393 +author: John Snow Labs +name: question_model_t5_tokenizer_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`question_model_t5_tokenizer_pipeline` is a English model originally trained by Mayank393. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/question_model_t5_tokenizer_pipeline_en_5.4.2_3.0_1724135538984.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/question_model_t5_tokenizer_pipeline_en_5.4.2_3.0_1724135538984.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("question_model_t5_tokenizer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("question_model_t5_tokenizer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|question_model_t5_tokenizer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.3 MB| + +## References + +https://huggingface.co/Mayank393/Question_Model_T5_Tokenizer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_en.md b/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_en.md new file mode 100644 index 00000000000000..4335069d2415a3 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English r1_coedit T5Transformer from owanr +author: John Snow Labs +name: r1_coedit +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`r1_coedit` is a English model originally trained by owanr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/r1_coedit_en_5.4.2_3.0_1724116112645.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/r1_coedit_en_5.4.2_3.0_1724116112645.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("r1_coedit","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("r1_coedit", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|r1_coedit| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/owanr/r1_coedit \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_pipeline_en.md new file mode 100644 index 00000000000000..c898cb1baa76e9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-r1_coedit_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English r1_coedit_pipeline pipeline T5Transformer from owanr +author: John Snow Labs +name: r1_coedit_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`r1_coedit_pipeline` is a English model originally trained by owanr. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/r1_coedit_pipeline_en_5.4.2_3.0_1724116262118.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/r1_coedit_pipeline_en_5.4.2_3.0_1724116262118.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("r1_coedit_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("r1_coedit_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|r1_coedit_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/owanr/r1_coedit + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_en.md new file mode 100644 index 00000000000000..b36e92d989b615 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English random_in_domain_5_demos_t5_small T5Transformer from Brendan +author: John Snow Labs +name: random_in_domain_5_demos_t5_small +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`random_in_domain_5_demos_t5_small` is a English model originally trained by Brendan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/random_in_domain_5_demos_t5_small_en_5.4.2_3.0_1724150538189.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/random_in_domain_5_demos_t5_small_en_5.4.2_3.0_1724150538189.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("random_in_domain_5_demos_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("random_in_domain_5_demos_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|random_in_domain_5_demos_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/Brendan/random-in-domain-5-demos-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..a22d708fc7a75e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-random_in_domain_5_demos_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English random_in_domain_5_demos_t5_small_pipeline pipeline T5Transformer from Brendan +author: John Snow Labs +name: random_in_domain_5_demos_t5_small_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`random_in_domain_5_demos_t5_small_pipeline` is a English model originally trained by Brendan. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/random_in_domain_5_demos_t5_small_pipeline_en_5.4.2_3.0_1724150555601.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/random_in_domain_5_demos_t5_small_pipeline_en_5.4.2_3.0_1724150555601.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("random_in_domain_5_demos_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("random_in_domain_5_demos_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|random_in_domain_5_demos_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/Brendan/random-in-domain-5-demos-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_en.md b/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_en.md new file mode 100644 index 00000000000000..dc7e6f7164062e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English real_prompt_100_500sync_all_gen_t5_small T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_100_500sync_all_gen_t5_small +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_100_500sync_all_gen_t5_small` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_100_500sync_all_gen_t5_small_en_5.4.2_3.0_1724135215312.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_100_500sync_all_gen_t5_small_en_5.4.2_3.0_1724135215312.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("real_prompt_100_500sync_all_gen_t5_small","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("real_prompt_100_500sync_all_gen_t5_small", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_100_500sync_all_gen_t5_small| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-100-500sync-all-gen-t5-small \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_pipeline_en.md new file mode 100644 index 00000000000000..f505af102d7490 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-real_prompt_100_500sync_all_gen_t5_small_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English real_prompt_100_500sync_all_gen_t5_small_pipeline pipeline T5Transformer from ShokSmile +author: John Snow Labs +name: real_prompt_100_500sync_all_gen_t5_small_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`real_prompt_100_500sync_all_gen_t5_small_pipeline` is a English model originally trained by ShokSmile. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/real_prompt_100_500sync_all_gen_t5_small_pipeline_en_5.4.2_3.0_1724135275130.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/real_prompt_100_500sync_all_gen_t5_small_pipeline_en_5.4.2_3.0_1724135275130.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("real_prompt_100_500sync_all_gen_t5_small_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("real_prompt_100_500sync_all_gen_t5_small_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|real_prompt_100_500sync_all_gen_t5_small_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/ShokSmile/real-prompt-100-500sync-all-gen-t5-small + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-retrained_en.md b/docs/_posts/ahmedlone127/2024-08-20-retrained_en.md new file mode 100644 index 00000000000000..94a345d3762863 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-retrained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English retrained T5Transformer from andreaolmos1990 +author: John Snow Labs +name: retrained +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`retrained` is a English model originally trained by andreaolmos1990. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/retrained_en_5.4.2_3.0_1724129737081.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/retrained_en_5.4.2_3.0_1724129737081.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("retrained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("retrained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|retrained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/andreaolmos1990/retrained \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-retrained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-retrained_pipeline_en.md new file mode 100644 index 00000000000000..a5458eafb0a978 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-retrained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English retrained_pipeline pipeline T5Transformer from andreaolmos1990 +author: John Snow Labs +name: retrained_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`retrained_pipeline` is a English model originally trained by andreaolmos1990. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/retrained_pipeline_en_5.4.2_3.0_1724129823942.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/retrained_pipeline_en_5.4.2_3.0_1724129823942.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("retrained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("retrained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|retrained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.3 GB| + +## References + +https://huggingface.co/andreaolmos1990/retrained + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_en.md b/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_en.md new file mode 100644 index 00000000000000..ff163864c5f79c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English russian_t5_comments_sum T5Transformer from ArtemYerokhin +author: John Snow Labs +name: russian_t5_comments_sum +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`russian_t5_comments_sum` is a English model originally trained by ArtemYerokhin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/russian_t5_comments_sum_en_5.4.2_3.0_1724138667035.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/russian_t5_comments_sum_en_5.4.2_3.0_1724138667035.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("russian_t5_comments_sum","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("russian_t5_comments_sum", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|russian_t5_comments_sum| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|996.8 MB| + +## References + +https://huggingface.co/ArtemYerokhin/ru_t5_comments_sum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_pipeline_en.md new file mode 100644 index 00000000000000..3ec0c5c9f28ec1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-russian_t5_comments_sum_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English russian_t5_comments_sum_pipeline pipeline T5Transformer from ArtemYerokhin +author: John Snow Labs +name: russian_t5_comments_sum_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`russian_t5_comments_sum_pipeline` is a English model originally trained by ArtemYerokhin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/russian_t5_comments_sum_pipeline_en_5.4.2_3.0_1724138713344.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/russian_t5_comments_sum_pipeline_en_5.4.2_3.0_1724138713344.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("russian_t5_comments_sum_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("russian_t5_comments_sum_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|russian_t5_comments_sum_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|996.8 MB| + +## References + +https://huggingface.co/ArtemYerokhin/ru_t5_comments_sum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_pipeline_ru.md b/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_pipeline_ru.md new file mode 100644 index 00000000000000..6704e28657a37e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_pipeline_ru.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Russian rut5_neuro_chgk_answering_pipeline pipeline T5Transformer from mary905el +author: John Snow Labs +name: rut5_neuro_chgk_answering_pipeline +date: 2024-08-20 +tags: [ru, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_neuro_chgk_answering_pipeline` is a Russian model originally trained by mary905el. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_neuro_chgk_answering_pipeline_ru_5.4.2_3.0_1724115091811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_neuro_chgk_answering_pipeline_ru_5.4.2_3.0_1724115091811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("rut5_neuro_chgk_answering_pipeline", lang = "ru") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("rut5_neuro_chgk_answering_pipeline", lang = "ru") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_neuro_chgk_answering_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|ru| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mary905el/ruT5_neuro_chgk_answering + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_ru.md b/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_ru.md new file mode 100644 index 00000000000000..3151906bed9d79 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-rut5_neuro_chgk_answering_ru.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Russian rut5_neuro_chgk_answering T5Transformer from mary905el +author: John Snow Labs +name: rut5_neuro_chgk_answering +date: 2024-08-20 +tags: [ru, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: ru +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`rut5_neuro_chgk_answering` is a Russian model originally trained by mary905el. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/rut5_neuro_chgk_answering_ru_5.4.2_3.0_1724115044479.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/rut5_neuro_chgk_answering_ru_5.4.2_3.0_1724115044479.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("rut5_neuro_chgk_answering","ru") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("rut5_neuro_chgk_answering", "ru") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|rut5_neuro_chgk_answering| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|ru| +|Size:|1.0 GB| + +## References + +https://huggingface.co/mary905el/ruT5_neuro_chgk_answering \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_en.md b/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_en.md new file mode 100644 index 00000000000000..bd48f1b281ef16 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc T5Transformer from Adarsh +author: John Snow Labs +name: scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc` is a English model originally trained by Adarsh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_en_5.4.2_3.0_1724129444826.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_en_5.4.2_3.0_1724129444826.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|933.4 MB| + +## References + +https://huggingface.co/Adarsh/SciFive-base-Pubmed_PMC-finetuned-SciFive-base-Pubmed-PMC \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline_en.md new file mode 100644 index 00000000000000..c989f93beed1b7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline pipeline T5Transformer from Adarsh +author: John Snow Labs +name: scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline` is a English model originally trained by Adarsh. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline_en_5.4.2_3.0_1724129508235.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline_en_5.4.2_3.0_1724129508235.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|scifive_base_pubmed_pmc_finetuned_scifive_base_pubmed_pmc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|933.4 MB| + +## References + +https://huggingface.co/Adarsh/SciFive-base-Pubmed_PMC-finetuned-SciFive-base-Pubmed-PMC + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_en.md b/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_en.md new file mode 100644 index 00000000000000..ba7fe2becff598 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1 T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_en_5.4.2_3.0_1724127946717.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_en_5.4.2_3.0_1724127946717.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|316.7 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline_en.md new file mode 100644 index 00000000000000..ba80a6ae1c25d0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline pipeline T5Transformer from sheoran95 +author: John Snow Labs +name: shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline` is a English model originally trained by sheoran95. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline_en_5.4.2_3.0_1724127964880.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline_en_5.4.2_3.0_1724127964880.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|shuffled_nodes_augmented_graphs_without_edge_document_level_t5_run1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|316.7 MB| + +## References + +https://huggingface.co/sheoran95/shuffled_nodes_augmented_graphs_without_edge_document_level_T5_run1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_pipeline_xx.md b/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_pipeline_xx.md new file mode 100644 index 00000000000000..bb9c302b118961 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_pipeline_xx.md @@ -0,0 +1,69 @@ +--- +layout: model +title: Multilingual slavic_t5_base_pipeline pipeline T5Transformer from TUKE-KEMT +author: John Snow Labs +name: slavic_t5_base_pipeline +date: 2024-08-20 +tags: [xx, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`slavic_t5_base_pipeline` is a Multilingual model originally trained by TUKE-KEMT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/slavic_t5_base_pipeline_xx_5.4.2_3.0_1724133986914.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/slavic_t5_base_pipeline_xx_5.4.2_3.0_1724133986914.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("slavic_t5_base_pipeline", lang = "xx") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("slavic_t5_base_pipeline", lang = "xx") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|slavic_t5_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|xx| +|Size:|1.8 GB| + +## References + +https://huggingface.co/TUKE-KEMT/slavic-t5-base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_xx.md b/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_xx.md new file mode 100644 index 00000000000000..725f93d344e891 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-slavic_t5_base_xx.md @@ -0,0 +1,86 @@ +--- +layout: model +title: Multilingual slavic_t5_base T5Transformer from TUKE-KEMT +author: John Snow Labs +name: slavic_t5_base +date: 2024-08-20 +tags: [xx, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: xx +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`slavic_t5_base` is a Multilingual model originally trained by TUKE-KEMT. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/slavic_t5_base_xx_5.4.2_3.0_1724133908149.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/slavic_t5_base_xx_5.4.2_3.0_1724133908149.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("slavic_t5_base","xx") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("slavic_t5_base", "xx") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|slavic_t5_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|xx| +|Size:|1.8 GB| + +## References + +https://huggingface.co/TUKE-KEMT/slavic-t5-base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_en.md b/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_en.md new file mode 100644 index 00000000000000..ed02cfca9c5f94 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English super_cool_model T5Transformer from lariskelmer +author: John Snow Labs +name: super_cool_model +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`super_cool_model` is a English model originally trained by lariskelmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/super_cool_model_en_5.4.2_3.0_1724125153105.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/super_cool_model_en_5.4.2_3.0_1724125153105.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("super_cool_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("super_cool_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|super_cool_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/lariskelmer/super-cool-model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_pipeline_en.md new file mode 100644 index 00000000000000..6ae4e18a17368c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-super_cool_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English super_cool_model_pipeline pipeline T5Transformer from lariskelmer +author: John Snow Labs +name: super_cool_model_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`super_cool_model_pipeline` is a English model originally trained by lariskelmer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/super_cool_model_pipeline_en_5.4.2_3.0_1724125212850.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/super_cool_model_pipeline_en_5.4.2_3.0_1724125212850.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("super_cool_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("super_cool_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|super_cool_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|179.1 MB| + +## References + +https://huggingface.co/lariskelmer/super-cool-model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_en.md b/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_en.md new file mode 100644 index 00000000000000..990e9c03f73321 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English synthea_t5_summarization_model T5Transformer from abymmathew +author: John Snow Labs +name: synthea_t5_summarization_model +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`synthea_t5_summarization_model` is a English model originally trained by abymmathew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/synthea_t5_summarization_model_en_5.4.2_3.0_1724118433173.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/synthea_t5_summarization_model_en_5.4.2_3.0_1724118433173.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("synthea_t5_summarization_model","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("synthea_t5_summarization_model", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|synthea_t5_summarization_model| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|317.4 MB| + +## References + +https://huggingface.co/abymmathew/synthea_t5_summarization_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_pipeline_en.md new file mode 100644 index 00000000000000..895c44ce7a9dc9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-synthea_t5_summarization_model_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English synthea_t5_summarization_model_pipeline pipeline T5Transformer from abymmathew +author: John Snow Labs +name: synthea_t5_summarization_model_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`synthea_t5_summarization_model_pipeline` is a English model originally trained by abymmathew. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/synthea_t5_summarization_model_pipeline_en_5.4.2_3.0_1724118456140.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/synthea_t5_summarization_model_pipeline_en_5.4.2_3.0_1724118456140.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("synthea_t5_summarization_model_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("synthea_t5_summarization_model_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|synthea_t5_summarization_model_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|317.4 MB| + +## References + +https://huggingface.co/abymmathew/synthea_t5_summarization_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_1_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_1_en.md new file mode 100644 index 00000000000000..0414348279bbb7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_1_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_1 T5Transformer from akira1608 +author: John Snow Labs +name: t5_1 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_1` is a English model originally trained by akira1608. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_1_en_5.4.2_3.0_1724115033791.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_1_en_5.4.2_3.0_1724115033791.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_1","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_1", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_1| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/akira1608/T5-1 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_1_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_1_pipeline_en.md new file mode 100644 index 00000000000000..63446578a10e5f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_1_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_1_pipeline pipeline T5Transformer from akira1608 +author: John Snow Labs +name: t5_1_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_1_pipeline` is a English model originally trained by akira1608. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_1_pipeline_en_5.4.2_3.0_1724115080426.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_1_pipeline_en_5.4.2_3.0_1724115080426.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_1_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_1_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_1_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/akira1608/T5-1 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_4m_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_4m_en.md new file mode 100644 index 00000000000000..01acbaafbc480f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_4m_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_4m T5Transformer from versae +author: John Snow Labs +name: t5_4m +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_4m` is a English model originally trained by versae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_4m_en_5.4.2_3.0_1724147859413.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_4m_en_5.4.2_3.0_1724147859413.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_4m","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_4m", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_4m| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/versae/t5-4m \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_4m_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_4m_pipeline_en.md new file mode 100644 index 00000000000000..bfc2a06e146e98 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_4m_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_4m_pipeline pipeline T5Transformer from versae +author: John Snow Labs +name: t5_4m_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_4m_pipeline` is a English model originally trained by versae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_4m_pipeline_en_5.4.2_3.0_1724147909748.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_4m_pipeline_en_5.4.2_3.0_1724147909748.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_4m_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_4m_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_4m_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/versae/t5-4m + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_en.md new file mode 100644 index 00000000000000..2c0716741635aa --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2012_2 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2012_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2012_2` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_2_en_5.4.2_3.0_1724151023498.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_2_en_5.4.2_3.0_1724151023498.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2012_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2012_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2012_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2012-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_pipeline_en.md new file mode 100644 index 00000000000000..189ddcda3bb8fc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2012_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2012_2_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2012_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2012_2_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_2_pipeline_en_5.4.2_3.0_1724151040331.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2012_2_pipeline_en_5.4.2_3.0_1724151040331.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2012_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2012_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2012_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2012-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_en.md new file mode 100644 index 00000000000000..5a20e34c1e599e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2015_0 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2015_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2015_0` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_0_en_5.4.2_3.0_1724134623739.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_0_en_5.4.2_3.0_1724134623739.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2015_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2015_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2015_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2015-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_pipeline_en.md new file mode 100644 index 00000000000000..af70d6ec84c59d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2015_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2015_0_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2015_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2015_0_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_0_pipeline_en_5.4.2_3.0_1724134640363.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2015_0_pipeline_en_5.4.2_3.0_1724134640363.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2015_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2015_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2015_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2015-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_en.md new file mode 100644 index 00000000000000..ad01f637ea0d4c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2017_7 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2017_7 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2017_7` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_7_en_5.4.2_3.0_1724121172624.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_7_en_5.4.2_3.0_1724121172624.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2017_7","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2017_7", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2017_7| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2017-7 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_pipeline_en.md new file mode 100644 index 00000000000000..b29736a01ca34e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2017_7_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2017_7_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2017_7_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2017_7_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_7_pipeline_en_5.4.2_3.0_1724121189532.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2017_7_pipeline_en_5.4.2_3.0_1724121189532.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2017_7_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2017_7_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2017_7_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.1 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2017-7 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_en.md new file mode 100644 index 00000000000000..a5501b9c8b80b7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_10 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_10 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_10` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_10_en_5.4.2_3.0_1724138920323.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_10_en_5.4.2_3.0_1724138920323.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_lm_wmt_2018_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_pipeline_en.md new file mode 100644 index 00000000000000..75485b9bf2c2b6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_lm_wmt_2018_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_lm_wmt_2018_10_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_lm_wmt_2018_10_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_lm_wmt_2018_10_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_10_pipeline_en_5.4.2_3.0_1724138940594.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_lm_wmt_2018_10_pipeline_en_5.4.2_3.0_1724138940594.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_lm_wmt_2018_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_lm_wmt_2018_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_lm_wmt_2018_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-lm-wmt-2018-10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_en.md new file mode 100644 index 00000000000000..79a9fcd48229f1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_5 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_5` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_5_en_5.4.2_3.0_1724129501249.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_5_en_5.4.2_3.0_1724129501249.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2017_5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|302.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_pipeline_en.md new file mode 100644 index 00000000000000..071b10821f8c6e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2017_5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2017_5_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2017_5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2017_5_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_5_pipeline_en_5.4.2_3.0_1724129529275.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2017_5_pipeline_en_5.4.2_3.0_1724129529275.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2017_5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2017_5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2017_5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|302.2 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2017-5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_en.md new file mode 100644 index 00000000000000..d27eb66fa355e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_8 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_8 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_8` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_8_en_5.4.2_3.0_1724113522373.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_8_en_5.4.2_3.0_1724113522373.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_8","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2018_8", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_8| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|299.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-8 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_pipeline_en.md new file mode 100644 index 00000000000000..69dbff71b3279b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2018_8_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2018_8_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2018_8_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2018_8_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_8_pipeline_en_5.4.2_3.0_1724113551354.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2018_8_pipeline_en_5.4.2_3.0_1724113551354.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2018_8_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2018_8_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2018_8_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|299.6 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2018-8 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_en.md new file mode 100644 index 00000000000000..544b57d59ebd04 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_10 T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_10 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_10` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_10_en_5.4.2_3.0_1724128032015.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_10_en_5.4.2_3.0_1724128032015.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_10","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_60m_poli_aff_2020_10", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_10| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|302.7 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-10 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_pipeline_en.md new file mode 100644 index 00000000000000..53a4215254aa93 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_60m_poli_aff_2020_10_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_60m_poli_aff_2020_10_pipeline pipeline T5Transformer from KaiNylund +author: John Snow Labs +name: t5_60m_poli_aff_2020_10_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_60m_poli_aff_2020_10_pipeline` is a English model originally trained by KaiNylund. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_10_pipeline_en_5.4.2_3.0_1724128060195.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_60m_poli_aff_2020_10_pipeline_en_5.4.2_3.0_1724128060195.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_60m_poli_aff_2020_10_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_60m_poli_aff_2020_10_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_60m_poli_aff_2020_10_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|302.7 MB| + +## References + +https://huggingface.co/KaiNylund/t5-60M-poli_aff-2020-10 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_9m_large_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_9m_large_en.md new file mode 100644 index 00000000000000..772274518a85a6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_9m_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_9m_large T5Transformer from versae +author: John Snow Labs +name: t5_9m_large +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_9m_large` is a English model originally trained by versae. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_9m_large_en_5.4.2_3.0_1724119802246.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_9m_large_en_5.4.2_3.0_1724119802246.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_9m_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_9m_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_9m_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/versae/t5-9m-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_en.md new file mode 100644 index 00000000000000..fdf025ad3b0055 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt1_khanq_eduqg T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt1_khanq_eduqg +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt1_khanq_eduqg` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt1_khanq_eduqg_en_5.4.2_3.0_1724138358690.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt1_khanq_eduqg_en_5.4.2_3.0_1724138358690.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt1_khanq_eduqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt1_khanq_eduqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt1_khanq_eduqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt1-khanq-eduqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_pipeline_en.md new file mode 100644 index 00000000000000..ed8143f64e75ac --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt1_khanq_eduqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt1_khanq_eduqg_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt1_khanq_eduqg_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt1_khanq_eduqg_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt1_khanq_eduqg_pipeline_en_5.4.2_3.0_1724138404783.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt1_khanq_eduqg_pipeline_en_5.4.2_3.0_1724138404783.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt1_khanq_eduqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt1_khanq_eduqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt1_khanq_eduqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt1-khanq-eduqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_en.md new file mode 100644 index 00000000000000..0524d396faff40 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_bt2_khanq_eduqg T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt2_khanq_eduqg +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt2_khanq_eduqg` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_eduqg_en_5.4.2_3.0_1724135035806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_eduqg_en_5.4.2_3.0_1724135035806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_bt2_khanq_eduqg","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_bt2_khanq_eduqg", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt2_khanq_eduqg| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt2-khanq-eduqg \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_pipeline_en.md new file mode 100644 index 00000000000000..cedc0a6db60191 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_bt2_khanq_eduqg_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_bt2_khanq_eduqg_pipeline pipeline T5Transformer from xiaothung +author: John Snow Labs +name: t5_base_bt2_khanq_eduqg_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_bt2_khanq_eduqg_pipeline` is a English model originally trained by xiaothung. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_eduqg_pipeline_en_5.4.2_3.0_1724135084370.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_bt2_khanq_eduqg_pipeline_en_5.4.2_3.0_1724135084370.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_bt2_khanq_eduqg_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_bt2_khanq_eduqg_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_bt2_khanq_eduqg_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/xiaothung/t5-base-bt2-khanq-eduqg + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_en.md new file mode 100644 index 00000000000000..65739b1e999938 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_detox T5Transformer from ChrisZeng +author: John Snow Labs +name: t5_base_detox +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_detox` is a English model originally trained by ChrisZeng. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_detox_en_5.4.2_3.0_1724141163368.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_detox_en_5.4.2_3.0_1724141163368.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_detox","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_detox", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_detox| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ChrisZeng/t5-base-detox \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_pipeline_en.md new file mode 100644 index 00000000000000..161ee8cfbf4347 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_detox_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_detox_pipeline pipeline T5Transformer from ChrisZeng +author: John Snow Labs +name: t5_base_detox_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_detox_pipeline` is a English model originally trained by ChrisZeng. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_detox_pipeline_en_5.4.2_3.0_1724141212174.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_detox_pipeline_en_5.4.2_3.0_1724141212174.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_detox_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_detox_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_detox_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/ChrisZeng/t5-base-detox + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_en.md new file mode 100644 index 00000000000000..8b867b4919403f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_dialogsum_seed32 T5Transformer from PSW +author: John Snow Labs +name: t5_base_dialogsum_seed32 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_dialogsum_seed32` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_dialogsum_seed32_en_5.4.2_3.0_1724115777787.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_dialogsum_seed32_en_5.4.2_3.0_1724115777787.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_dialogsum_seed32","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_dialogsum_seed32", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_dialogsum_seed32| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-dialogsum-seed32 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_pipeline_en.md new file mode 100644 index 00000000000000..e478199fef5c04 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_dialogsum_seed32_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_dialogsum_seed32_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_dialogsum_seed32_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_dialogsum_seed32_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_dialogsum_seed32_pipeline_en_5.4.2_3.0_1724115840081.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_dialogsum_seed32_pipeline_en_5.4.2_3.0_1724115840081.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_dialogsum_seed32_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_dialogsum_seed32_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_dialogsum_seed32_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-dialogsum-seed32 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_en.md new file mode 100644 index 00000000000000..3ccbd8c7d175c8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1724133467146.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_en_5.4.2_3.0_1724133467146.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|945.3 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-128-finetuned-squad-infilling-seed-2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline_en.md new file mode 100644 index 00000000000000..05dfbc7e1749ed --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1724133530975.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline_en_5.4.2_3.0_1724133530975.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_few_shot_k_128_finetuned_squad_infilling_seed_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|945.3 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-base-few-shot-k-128-finetuned-squad-infilling-seed-2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_en.md new file mode 100644 index 00000000000000..be3e38feef245c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_g2e_translation T5Transformer from oren186 +author: John Snow Labs +name: t5_base_finetuned_g2e_translation +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_g2e_translation` is a English model originally trained by oren186. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_g2e_translation_en_5.4.2_3.0_1724114889239.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_g2e_translation_en_5.4.2_3.0_1724114889239.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_g2e_translation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_g2e_translation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_g2e_translation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/oren186/t5-base-finetuned-G2E-Translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_pipeline_en.md new file mode 100644 index 00000000000000..0f7645ab227928 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_g2e_translation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_g2e_translation_pipeline pipeline T5Transformer from oren186 +author: John Snow Labs +name: t5_base_finetuned_g2e_translation_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_g2e_translation_pipeline` is a English model originally trained by oren186. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_g2e_translation_pipeline_en_5.4.2_3.0_1724114935446.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_g2e_translation_pipeline_en_5.4.2_3.0_1724114935446.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_g2e_translation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_g2e_translation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_g2e_translation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/oren186/t5-base-finetuned-G2E-Translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_en.md new file mode 100644 index 00000000000000..696f53d7e79070 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_finetuned_qasc_sardinian T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_qasc_sardinian +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_qasc_sardinian` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_qasc_sardinian_en_5.4.2_3.0_1724122247612.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_qasc_sardinian_en_5.4.2_3.0_1724122247612.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_finetuned_qasc_sardinian","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_finetuned_qasc_sardinian", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_qasc_sardinian| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|932.0 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-qasc-sc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_pipeline_en.md new file mode 100644 index 00000000000000..00a5342bbb9727 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_finetuned_qasc_sardinian_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_finetuned_qasc_sardinian_pipeline pipeline T5Transformer from mrm8488 +author: John Snow Labs +name: t5_base_finetuned_qasc_sardinian_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_finetuned_qasc_sardinian_pipeline` is a English model originally trained by mrm8488. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_qasc_sardinian_pipeline_en_5.4.2_3.0_1724122318099.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_finetuned_qasc_sardinian_pipeline_en_5.4.2_3.0_1724122318099.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_finetuned_qasc_sardinian_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_finetuned_qasc_sardinian_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_finetuned_qasc_sardinian_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|932.0 MB| + +## References + +https://huggingface.co/mrm8488/t5-base-finetuned-qasc-sc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_en.md new file mode 100644 index 00000000000000..526294933995e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_fullwnc_5epoch2_2dc8dc72 T5Transformer from lindsayng +author: John Snow Labs +name: t5_base_fullwnc_5epoch2_2dc8dc72 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_fullwnc_5epoch2_2dc8dc72` is a English model originally trained by lindsayng. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_fullwnc_5epoch2_2dc8dc72_en_5.4.2_3.0_1724117148042.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_fullwnc_5epoch2_2dc8dc72_en_5.4.2_3.0_1724117148042.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_fullwnc_5epoch2_2dc8dc72","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_fullwnc_5epoch2_2dc8dc72", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_fullwnc_5epoch2_2dc8dc72| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lindsayng/t5-base-fullwnc-5epoch2-2dc8dc72 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_pipeline_en.md new file mode 100644 index 00000000000000..8e258aa399ba77 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_fullwnc_5epoch2_2dc8dc72_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_fullwnc_5epoch2_2dc8dc72_pipeline pipeline T5Transformer from lindsayng +author: John Snow Labs +name: t5_base_fullwnc_5epoch2_2dc8dc72_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_fullwnc_5epoch2_2dc8dc72_pipeline` is a English model originally trained by lindsayng. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_fullwnc_5epoch2_2dc8dc72_pipeline_en_5.4.2_3.0_1724117196885.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_fullwnc_5epoch2_2dc8dc72_pipeline_en_5.4.2_3.0_1724117196885.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_fullwnc_5epoch2_2dc8dc72_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_fullwnc_5epoch2_2dc8dc72_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_fullwnc_5epoch2_2dc8dc72_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/lindsayng/t5-base-fullwnc-5epoch2-2dc8dc72 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_en.md new file mode 100644 index 00000000000000..ba3bd489b1f24b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_samsum_seed42 T5Transformer from PSW +author: John Snow Labs +name: t5_base_samsum_seed42 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_samsum_seed42` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed42_en_5.4.2_3.0_1724139341392.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed42_en_5.4.2_3.0_1724139341392.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_samsum_seed42","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_samsum_seed42", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_samsum_seed42| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-samsum-seed42 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_pipeline_en.md new file mode 100644 index 00000000000000..ec89fa7a8756d9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_samsum_seed42_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_samsum_seed42_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_samsum_seed42_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_samsum_seed42_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed42_pipeline_en_5.4.2_3.0_1724139389404.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_samsum_seed42_pipeline_en_5.4.2_3.0_1724139389404.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_samsum_seed42_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_samsum_seed42_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_samsum_seed42_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/PSW/t5-base-samsum-seed42 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_en.md new file mode 100644 index 00000000000000..fbaaa654623e31 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sft_books T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_books +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_books` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_books_en_5.4.2_3.0_1724146248256.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_books_en_5.4.2_3.0_1724146248256.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sft_books","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sft_books", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_books| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-books \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_pipeline_en.md new file mode 100644 index 00000000000000..7d7d172ab3a8c0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_books_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sft_books_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_books_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_books_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_books_pipeline_en_5.4.2_3.0_1724146300703.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_books_pipeline_en_5.4.2_3.0_1724146300703.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sft_books_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sft_books_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_books_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-books + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_en.md new file mode 100644 index 00000000000000..00f843b41e7612 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_sft_instruments T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_instruments +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_instruments` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_instruments_en_5.4.2_3.0_1724121342422.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_instruments_en_5.4.2_3.0_1724121342422.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_sft_instruments","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_sft_instruments", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_instruments| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|953.2 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-instruments \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_pipeline_en.md new file mode 100644 index 00000000000000..bcc0f47c3d295b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_sft_instruments_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_sft_instruments_pipeline pipeline T5Transformer from yashonwu +author: John Snow Labs +name: t5_base_sft_instruments_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_sft_instruments_pipeline` is a English model originally trained by yashonwu. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_sft_instruments_pipeline_en_5.4.2_3.0_1724121406986.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_sft_instruments_pipeline_en_5.4.2_3.0_1724121406986.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_sft_instruments_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_sft_instruments_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_sft_instruments_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|953.2 MB| + +## References + +https://huggingface.co/yashonwu/t5-base-sft-instruments + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_en.md new file mode 100644 index 00000000000000..80da6d2a1217ff --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tedxjp_1front_1body_1rear T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_1front_1body_1rear +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_1front_1body_1rear` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1front_1body_1rear_en_5.4.2_3.0_1724114647779.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1front_1body_1rear_en_5.4.2_3.0_1724114647779.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tedxjp_1front_1body_1rear","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tedxjp_1front_1body_1rear", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_1front_1body_1rear| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-1front-1body-1rear \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_pipeline_en.md new file mode 100644 index 00000000000000..57819d346e2ef7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tedxjp_1front_1body_1rear_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tedxjp_1front_1body_1rear_pipeline pipeline T5Transformer from Padomin +author: John Snow Labs +name: t5_base_tedxjp_1front_1body_1rear_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tedxjp_1front_1body_1rear_pipeline` is a English model originally trained by Padomin. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1front_1body_1rear_pipeline_en_5.4.2_3.0_1724114698518.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tedxjp_1front_1body_1rear_pipeline_en_5.4.2_3.0_1724114698518.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tedxjp_1front_1body_1rear_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tedxjp_1front_1body_1rear_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tedxjp_1front_1body_1rear_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Padomin/t5-base-TEDxJP-1front-1body-1rear + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_en.md new file mode 100644 index 00000000000000..cd47c11340f832 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_base_tweetsumm_seed55 T5Transformer from PSW +author: John Snow Labs +name: t5_base_tweetsumm_seed55 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tweetsumm_seed55` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tweetsumm_seed55_en_5.4.2_3.0_1724125235462.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tweetsumm_seed55_en_5.4.2_3.0_1724125235462.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_base_tweetsumm_seed55","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_base_tweetsumm_seed55", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tweetsumm_seed55| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.6 MB| + +## References + +https://huggingface.co/PSW/t5-base-tweetsumm-seed55 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_pipeline_en.md new file mode 100644 index 00000000000000..afdf3e206933e2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_base_tweetsumm_seed55_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_base_tweetsumm_seed55_pipeline pipeline T5Transformer from PSW +author: John Snow Labs +name: t5_base_tweetsumm_seed55_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_base_tweetsumm_seed55_pipeline` is a English model originally trained by PSW. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_base_tweetsumm_seed55_pipeline_en_5.4.2_3.0_1724125285034.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_base_tweetsumm_seed55_pipeline_en_5.4.2_3.0_1724125285034.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_base_tweetsumm_seed55_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_base_tweetsumm_seed55_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_base_tweetsumm_seed55_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.6 MB| + +## References + +https://huggingface.co/PSW/t5-base-tweetsumm-seed55 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_en.md new file mode 100644 index 00000000000000..2a9885f37a05cc --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_for_translation_roval15 T5Transformer from roval15 +author: John Snow Labs +name: t5_for_translation_roval15 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_for_translation_roval15` is a English model originally trained by roval15. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_for_translation_roval15_en_5.4.2_3.0_1724131931492.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_for_translation_roval15_en_5.4.2_3.0_1724131931492.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_for_translation_roval15","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_for_translation_roval15", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_for_translation_roval15| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.4 MB| + +## References + +https://huggingface.co/roval15/t5_for_translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_pipeline_en.md new file mode 100644 index 00000000000000..89f6231161c06c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_for_translation_roval15_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_for_translation_roval15_pipeline pipeline T5Transformer from roval15 +author: John Snow Labs +name: t5_for_translation_roval15_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_for_translation_roval15_pipeline` is a English model originally trained by roval15. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_for_translation_roval15_pipeline_en_5.4.2_3.0_1724131934086.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_for_translation_roval15_pipeline_en_5.4.2_3.0_1724131934086.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_for_translation_roval15_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_for_translation_roval15_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_for_translation_roval15_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.4 MB| + +## References + +https://huggingface.co/roval15/t5_for_translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_en.md new file mode 100644 index 00000000000000..29f8350ca5faa7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_esnli_impli_figurative T5Transformer from Rachneet +author: John Snow Labs +name: t5_large_esnli_impli_figurative +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_esnli_impli_figurative` is a English model originally trained by Rachneet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_esnli_impli_figurative_en_5.4.2_3.0_1724113148029.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_esnli_impli_figurative_en_5.4.2_3.0_1724113148029.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_esnli_impli_figurative","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_esnli_impli_figurative", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_esnli_impli_figurative| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Rachneet/T5-large-esnli-impli-figurative \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_pipeline_en.md new file mode 100644 index 00000000000000..34a60d49308973 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_esnli_impli_figurative_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_esnli_impli_figurative_pipeline pipeline T5Transformer from Rachneet +author: John Snow Labs +name: t5_large_esnli_impli_figurative_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_esnli_impli_figurative_pipeline` is a English model originally trained by Rachneet. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_esnli_impli_figurative_pipeline_en_5.4.2_3.0_1724113291737.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_esnli_impli_figurative_pipeline_en_5.4.2_3.0_1724113291737.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_esnli_impli_figurative_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_esnli_impli_figurative_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_esnli_impli_figurative_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Rachneet/T5-large-esnli-impli-figurative + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_en.md new file mode 100644 index 00000000000000..4f8ee01082e3e0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_laschulz T5Transformer from laschulz +author: John Snow Labs +name: t5_large_laschulz +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_laschulz` is a English model originally trained by laschulz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_laschulz_en_5.4.2_3.0_1724115966455.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_laschulz_en_5.4.2_3.0_1724115966455.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_laschulz","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_laschulz", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_laschulz| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/laschulz/t5-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_pipeline_en.md new file mode 100644 index 00000000000000..097a4e320f7938 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_laschulz_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_laschulz_pipeline pipeline T5Transformer from laschulz +author: John Snow Labs +name: t5_large_laschulz_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_laschulz_pipeline` is a English model originally trained by laschulz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_laschulz_pipeline_en_5.4.2_3.0_1724116136534.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_laschulz_pipeline_en_5.4.2_3.0_1724116136534.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_laschulz_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_laschulz_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_laschulz_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/laschulz/t5-large + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_en.md new file mode 100644 index 00000000000000..e8ad9d2f501f52 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_mod_translation T5Transformer from guyhadad01 +author: John Snow Labs +name: t5_large_mod_translation +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_mod_translation` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_mod_translation_en_5.4.2_3.0_1724141893714.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_mod_translation_en_5.4.2_3.0_1724141893714.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_mod_translation","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_mod_translation", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_mod_translation| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/guyhadad01/t5-large-mod-translation \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_pipeline_en.md new file mode 100644 index 00000000000000..d630d96dc3429d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_mod_translation_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_mod_translation_pipeline pipeline T5Transformer from guyhadad01 +author: John Snow Labs +name: t5_large_mod_translation_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_mod_translation_pipeline` is a English model originally trained by guyhadad01. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_mod_translation_pipeline_en_5.4.2_3.0_1724142031590.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_mod_translation_pipeline_en_5.4.2_3.0_1724142031590.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_mod_translation_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_mod_translation_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_mod_translation_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/guyhadad01/t5-large-mod-translation + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_en.md new file mode 100644 index 00000000000000..b8e494c42739f6 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_large_quartz T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_large_quartz +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_quartz` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_quartz_en_5.4.2_3.0_1724124591693.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_quartz_en_5.4.2_3.0_1724124591693.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_large_quartz","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_large_quartz", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_quartz| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/SeongwooKim/T5-large-quartz \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_pipeline_en.md new file mode 100644 index 00000000000000..0e03686387ce63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_large_quartz_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_large_quartz_pipeline pipeline T5Transformer from SeongwooKim +author: John Snow Labs +name: t5_large_quartz_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_large_quartz_pipeline` is a English model originally trained by SeongwooKim. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_large_quartz_pipeline_en_5.4.2_3.0_1724124743754.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_large_quartz_pipeline_en_5.4.2_3.0_1724124743754.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_large_quartz_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_large_quartz_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_large_quartz_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|2.9 GB| + +## References + +https://huggingface.co/SeongwooKim/T5-large-quartz + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_en.md new file mode 100644 index 00000000000000..4318050e7e5ac8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_qg_model_with_answer2 T5Transformer from santhoshkolloju +author: John Snow Labs +name: t5_qg_model_with_answer2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_qg_model_with_answer2` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_qg_model_with_answer2_en_5.4.2_3.0_1724150670664.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_qg_model_with_answer2_en_5.4.2_3.0_1724150670664.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_qg_model_with_answer2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_qg_model_with_answer2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_qg_model_with_answer2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|985.9 MB| + +## References + +https://huggingface.co/santhoshkolloju/t5_qg_model_with_answer2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_pipeline_en.md new file mode 100644 index 00000000000000..acc484e6ed858e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_qg_model_with_answer2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_qg_model_with_answer2_pipeline pipeline T5Transformer from santhoshkolloju +author: John Snow Labs +name: t5_qg_model_with_answer2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_qg_model_with_answer2_pipeline` is a English model originally trained by santhoshkolloju. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_qg_model_with_answer2_pipeline_en_5.4.2_3.0_1724150725055.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_qg_model_with_answer2_pipeline_en_5.4.2_3.0_1724150725055.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_qg_model_with_answer2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_qg_model_with_answer2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_qg_model_with_answer2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|985.9 MB| + +## References + +https://huggingface.co/santhoshkolloju/t5_qg_model_with_answer2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_en.md new file mode 100644 index 00000000000000..76e2182d8b7c4c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_russian_finetuned T5Transformer from KabanovDenis +author: John Snow Labs +name: t5_russian_finetuned +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_russian_finetuned` is a English model originally trained by KabanovDenis. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_russian_finetuned_en_5.4.2_3.0_1724136279768.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_russian_finetuned_en_5.4.2_3.0_1724136279768.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_russian_finetuned","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_russian_finetuned", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_russian_finetuned| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|991.4 MB| + +## References + +https://huggingface.co/KabanovDenis/t5_ru_finetuned \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_pipeline_en.md new file mode 100644 index 00000000000000..988f5f3906b9ce --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_russian_finetuned_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_russian_finetuned_pipeline pipeline T5Transformer from KabanovDenis +author: John Snow Labs +name: t5_russian_finetuned_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_russian_finetuned_pipeline` is a English model originally trained by KabanovDenis. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_russian_finetuned_pipeline_en_5.4.2_3.0_1724136333256.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_russian_finetuned_pipeline_en_5.4.2_3.0_1724136333256.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_russian_finetuned_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_russian_finetuned_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_russian_finetuned_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|991.4 MB| + +## References + +https://huggingface.co/KabanovDenis/t5_ru_finetuned + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_s19_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_s19_en.md new file mode 100644 index 00000000000000..25bc3a5be09719 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_s19_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_s19 T5Transformer from alraisi +author: John Snow Labs +name: t5_s19 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_s19` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_s19_en_5.4.2_3.0_1724117472200.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_s19_en_5.4.2_3.0_1724117472200.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_s19","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_s19", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_s19| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.9 MB| + +## References + +https://huggingface.co/alraisi/t5-s19 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_s19_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_s19_pipeline_en.md new file mode 100644 index 00000000000000..9a16e1e3eca125 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_s19_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_s19_pipeline pipeline T5Transformer from alraisi +author: John Snow Labs +name: t5_s19_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_s19_pipeline` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_s19_pipeline_en_5.4.2_3.0_1724117489633.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_s19_pipeline_en_5.4.2_3.0_1724117489633.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_s19_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_s19_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_s19_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.9 MB| + +## References + +https://huggingface.co/alraisi/t5-s19 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_en.md new file mode 100644 index 00000000000000..c8ff48cdfd2b77 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_slovene_large_oasst_pairs T5Transformer from vh-student +author: John Snow Labs +name: t5_slovene_large_oasst_pairs +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_slovene_large_oasst_pairs` is a English model originally trained by vh-student. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_pairs_en_5.4.2_3.0_1724138751930.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_pairs_en_5.4.2_3.0_1724138751930.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_slovene_large_oasst_pairs","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_slovene_large_oasst_pairs", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_slovene_large_oasst_pairs| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/vh-student/t5-sl-large-oasst-pairs \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_pipeline_en.md new file mode 100644 index 00000000000000..46259003121a9f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_slovene_large_oasst_pairs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_slovene_large_oasst_pairs_pipeline pipeline T5Transformer from vh-student +author: John Snow Labs +name: t5_slovene_large_oasst_pairs_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_slovene_large_oasst_pairs_pipeline` is a English model originally trained by vh-student. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_pairs_pipeline_en_5.4.2_3.0_1724138895064.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_slovene_large_oasst_pairs_pipeline_en_5.4.2_3.0_1724138895064.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_slovene_large_oasst_pairs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_slovene_large_oasst_pairs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_slovene_large_oasst_pairs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/vh-student/t5-sl-large-oasst-pairs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_en.md new file mode 100644 index 00000000000000..7070292b53f780 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_1024 T5Transformer from bangnbx +author: John Snow Labs +name: t5_small_1024 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_1024` is a English model originally trained by bangnbx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_1024_en_5.4.2_3.0_1724120293571.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_1024_en_5.4.2_3.0_1724120293571.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_1024","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_1024", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_1024| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/bangnbx/t5-small-1024 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_pipeline_en.md new file mode 100644 index 00000000000000..f8ded6e80bbc61 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_1024_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_1024_pipeline pipeline T5Transformer from bangnbx +author: John Snow Labs +name: t5_small_1024_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_1024_pipeline` is a English model originally trained by bangnbx. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_1024_pipeline_en_5.4.2_3.0_1724120310444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_1024_pipeline_en_5.4.2_3.0_1724120310444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_1024_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_1024_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_1024_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|349.7 MB| + +## References + +https://huggingface.co/bangnbx/t5-small-1024 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_en.md new file mode 100644 index 00000000000000..afd367acb2eb59 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_few_shot_k_16_finetuned_squad_seed_0 T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_16_finetuned_squad_seed_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_16_finetuned_squad_seed_0` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_0_en_5.4.2_3.0_1724129288498.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_0_en_5.4.2_3.0_1724129288498.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_few_shot_k_16_finetuned_squad_seed_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_few_shot_k_16_finetuned_squad_seed_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_16_finetuned_squad_seed_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|287.4 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-16-finetuned-squad-seed-0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline_en.md new file mode 100644 index 00000000000000..9c3e15df8c7ea7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline pipeline T5Transformer from anas-awadalla +author: John Snow Labs +name: t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline` is a English model originally trained by anas-awadalla. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline_en_5.4.2_3.0_1724129323508.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline_en_5.4.2_3.0_1724129323508.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_few_shot_k_16_finetuned_squad_seed_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|287.4 MB| + +## References + +https://huggingface.co/anas-awadalla/t5-small-few-shot-k-16-finetuned-squad-seed-0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_en.md new file mode 100644 index 00000000000000..9080c0cd64e11d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned16_english_tonga_tonga_islands_german T5Transformer from alexrfelicio +author: John Snow Labs +name: t5_small_finetuned16_english_tonga_tonga_islands_german +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned16_english_tonga_tonga_islands_german` is a English model originally trained by alexrfelicio. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned16_english_tonga_tonga_islands_german_en_5.4.2_3.0_1724133040513.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned16_english_tonga_tonga_islands_german_en_5.4.2_3.0_1724133040513.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned16_english_tonga_tonga_islands_german","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned16_english_tonga_tonga_islands_german", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned16_english_tonga_tonga_islands_german| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|308.7 MB| + +## References + +https://huggingface.co/alexrfelicio/t5-small-finetuned16-en-to-de \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline_en.md new file mode 100644 index 00000000000000..b1b997727cd0bf --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline pipeline T5Transformer from alexrfelicio +author: John Snow Labs +name: t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline` is a English model originally trained by alexrfelicio. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline_en_5.4.2_3.0_1724133061811.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline_en_5.4.2_3.0_1724133061811.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned16_english_tonga_tonga_islands_german_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|308.7 MB| + +## References + +https://huggingface.co/alexrfelicio/t5-small-finetuned16-en-to-de + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_en.md new file mode 100644 index 00000000000000..486fbf401833e7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_bias_72bc782c T5Transformer from ajsmith201 +author: John Snow Labs +name: t5_small_finetuned_bias_72bc782c +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_bias_72bc782c` is a English model originally trained by ajsmith201. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bias_72bc782c_en_5.4.2_3.0_1724121301190.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bias_72bc782c_en_5.4.2_3.0_1724121301190.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_bias_72bc782c","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_bias_72bc782c", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_bias_72bc782c| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|339.3 MB| + +## References + +https://huggingface.co/ajsmith201/t5-small-finetuned-bias-72bc782c \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_pipeline_en.md new file mode 100644 index 00000000000000..8cf34a2aef4fca --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_bias_72bc782c_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_bias_72bc782c_pipeline pipeline T5Transformer from ajsmith201 +author: John Snow Labs +name: t5_small_finetuned_bias_72bc782c_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_bias_72bc782c_pipeline` is a English model originally trained by ajsmith201. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bias_72bc782c_pipeline_en_5.4.2_3.0_1724121320444.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_bias_72bc782c_pipeline_en_5.4.2_3.0_1724121320444.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_bias_72bc782c_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_bias_72bc782c_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_bias_72bc782c_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|339.3 MB| + +## References + +https://huggingface.co/ajsmith201/t5-small-finetuned-bias-72bc782c + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_en.md new file mode 100644 index 00000000000000..a27118184b35e5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_med T5Transformer from alraisi +author: John Snow Labs +name: t5_small_finetuned_med +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_med` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_med_en_5.4.2_3.0_1724124931139.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_med_en_5.4.2_3.0_1724124931139.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_med","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_med", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_med| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|329.9 MB| + +## References + +https://huggingface.co/alraisi/t5-small-finetuned-med \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_pipeline_en.md new file mode 100644 index 00000000000000..dbfb7f84477a68 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_med_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_med_pipeline pipeline T5Transformer from alraisi +author: John Snow Labs +name: t5_small_finetuned_med_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_med_pipeline` is a English model originally trained by alraisi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_med_pipeline_en_5.4.2_3.0_1724124951453.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_med_pipeline_en_5.4.2_3.0_1724124951453.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_med_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_med_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_med_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|329.9 MB| + +## References + +https://huggingface.co/alraisi/t5-small-finetuned-med + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_en.md new file mode 100644 index 00000000000000..f93f5ba244a5b4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_wikisql_averagenlpenjoyer T5Transformer from AverageNLPEnjoyer +author: John Snow Labs +name: t5_small_finetuned_wikisql_averagenlpenjoyer +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_wikisql_averagenlpenjoyer` is a English model originally trained by AverageNLPEnjoyer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_wikisql_averagenlpenjoyer_en_5.4.2_3.0_1724129206672.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_wikisql_averagenlpenjoyer_en_5.4.2_3.0_1724129206672.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_wikisql_averagenlpenjoyer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_wikisql_averagenlpenjoyer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_wikisql_averagenlpenjoyer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|342.4 MB| + +## References + +https://huggingface.co/AverageNLPEnjoyer/t5-small-finetuned-wikisql \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline_en.md new file mode 100644 index 00000000000000..0f24ffddcef49b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline pipeline T5Transformer from AverageNLPEnjoyer +author: John Snow Labs +name: t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline` is a English model originally trained by AverageNLPEnjoyer. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline_en_5.4.2_3.0_1724129225856.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline_en_5.4.2_3.0_1724129225856.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_wikisql_averagenlpenjoyer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|342.4 MB| + +## References + +https://huggingface.co/AverageNLPEnjoyer/t5-small-finetuned-wikisql + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_en.md new file mode 100644 index 00000000000000..9ae6e454f20da9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_hugginfacexf T5Transformer from hugginfacexf +author: John Snow Labs +name: t5_small_finetuned_xsum_hugginfacexf +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_hugginfacexf` is a English model originally trained by hugginfacexf. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_hugginfacexf_en_5.4.2_3.0_1724136195984.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_hugginfacexf_en_5.4.2_3.0_1724136195984.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_hugginfacexf","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_finetuned_xsum_hugginfacexf", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_hugginfacexf| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|337.8 MB| + +## References + +https://huggingface.co/hugginfacexf/t5-small-finetuned-xsum \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_pipeline_en.md new file mode 100644 index 00000000000000..d433f7a23d092f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_finetuned_xsum_hugginfacexf_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_finetuned_xsum_hugginfacexf_pipeline pipeline T5Transformer from hugginfacexf +author: John Snow Labs +name: t5_small_finetuned_xsum_hugginfacexf_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_finetuned_xsum_hugginfacexf_pipeline` is a English model originally trained by hugginfacexf. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_hugginfacexf_pipeline_en_5.4.2_3.0_1724136213312.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_finetuned_xsum_hugginfacexf_pipeline_en_5.4.2_3.0_1724136213312.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_finetuned_xsum_hugginfacexf_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_finetuned_xsum_hugginfacexf_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_finetuned_xsum_hugginfacexf_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|337.8 MB| + +## References + +https://huggingface.co/hugginfacexf/t5-small-finetuned-xsum + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_en.md new file mode 100644 index 00000000000000..b3de7d4a035103 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_ruleviewer T5Transformer from Tgratzi +author: John Snow Labs +name: t5_small_ruleviewer +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_ruleviewer` is a English model originally trained by Tgratzi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ruleviewer_en_5.4.2_3.0_1724119898556.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ruleviewer_en_5.4.2_3.0_1724119898556.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_ruleviewer","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_ruleviewer", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ruleviewer| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|345.0 MB| + +## References + +https://huggingface.co/Tgratzi/t5-small-ruleviewer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_pipeline_en.md new file mode 100644 index 00000000000000..06c4fe366817d7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_ruleviewer_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_ruleviewer_pipeline pipeline T5Transformer from Tgratzi +author: John Snow Labs +name: t5_small_ruleviewer_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_ruleviewer_pipeline` is a English model originally trained by Tgratzi. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_ruleviewer_pipeline_en_5.4.2_3.0_1724119916089.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_ruleviewer_pipeline_en_5.4.2_3.0_1724119916089.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_ruleviewer_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_ruleviewer_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_ruleviewer_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|345.0 MB| + +## References + +https://huggingface.co/Tgratzi/t5-small-ruleviewer + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_en.md new file mode 100644 index 00000000000000..0612f9c902fce2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_small_squad_newsqa_qag_trained T5Transformer from StellarMilk +author: John Snow Labs +name: t5_small_squad_newsqa_qag_trained +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_newsqa_qag_trained` is a English model originally trained by StellarMilk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_newsqa_qag_trained_en_5.4.2_3.0_1724138422175.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_newsqa_qag_trained_en_5.4.2_3.0_1724138422175.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_small_squad_newsqa_qag_trained","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_small_squad_newsqa_qag_trained", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_newsqa_qag_trained| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|341.3 MB| + +## References + +https://huggingface.co/StellarMilk/t5-small-squad-newsqa-qag-trained \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_pipeline_en.md new file mode 100644 index 00000000000000..909f036ff4e565 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_small_squad_newsqa_qag_trained_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_small_squad_newsqa_qag_trained_pipeline pipeline T5Transformer from StellarMilk +author: John Snow Labs +name: t5_small_squad_newsqa_qag_trained_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_small_squad_newsqa_qag_trained_pipeline` is a English model originally trained by StellarMilk. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_small_squad_newsqa_qag_trained_pipeline_en_5.4.2_3.0_1724138439615.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_small_squad_newsqa_qag_trained_pipeline_en_5.4.2_3.0_1724138439615.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_small_squad_newsqa_qag_trained_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_small_squad_newsqa_qag_trained_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_small_squad_newsqa_qag_trained_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|341.3 MB| + +## References + +https://huggingface.co/StellarMilk/t5-small-squad-newsqa-qag-trained + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_en.md new file mode 100644 index 00000000000000..5f1aac00a41f3e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da T5Transformer from diegor2 +author: John Snow Labs +name: t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da` is a English model originally trained by diegor2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_en_5.4.2_3.0_1724131715484.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_en_5.4.2_3.0_1724131715484.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|24.0 MB| + +## References + +https://huggingface.co/diegor2/t5-tiny-random-length-96-learning_rate-0.0001-weight_decay-0.01-finetu-truncated-5e15da \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline_en.md new file mode 100644 index 00000000000000..465206a4fa1b3f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline pipeline T5Transformer from diegor2 +author: John Snow Labs +name: t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline` is a English model originally trained by diegor2. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline_en_5.4.2_3.0_1724131717016.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline_en_5.4.2_3.0_1724131717016.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_tiny_random_length_96_learning_rate_0_0001_weight_decay_0_01_finetu_truncated_5e15da_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|24.0 MB| + +## References + +https://huggingface.co/diegor2/t5-tiny-random-length-96-learning_rate-0.0001-weight_decay-0.01-finetu-truncated-5e15da + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_en.md new file mode 100644 index 00000000000000..d72acb6709e3f9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5_ultradetox T5Transformer from slewie +author: John Snow Labs +name: t5_ultradetox +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ultradetox` is a English model originally trained by slewie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ultradetox_en_5.4.2_3.0_1724137129511.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ultradetox_en_5.4.2_3.0_1724137129511.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5_ultradetox","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5_ultradetox", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ultradetox| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/slewie/t5-ultradetox \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_pipeline_en.md new file mode 100644 index 00000000000000..943fb8c77d3d82 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5_ultradetox_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5_ultradetox_pipeline pipeline T5Transformer from slewie +author: John Snow Labs +name: t5_ultradetox_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5_ultradetox_pipeline` is a English model originally trained by slewie. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5_ultradetox_pipeline_en_5.4.2_3.0_1724137183674.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5_ultradetox_pipeline_en_5.4.2_3.0_1724137183674.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5_ultradetox_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5_ultradetox_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5_ultradetox_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/slewie/t5-ultradetox + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_en.md new file mode 100644 index 00000000000000..956599d551bbe2 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5doctalk T5Transformer from Elfsong +author: John Snow Labs +name: t5doctalk +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5doctalk` is a English model originally trained by Elfsong. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5doctalk_en_5.4.2_3.0_1724135813267.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5doctalk_en_5.4.2_3.0_1724135813267.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5doctalk","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5doctalk", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5doctalk| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Elfsong/t5doctalk \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_pipeline_en.md new file mode 100644 index 00000000000000..de30ac1f8c8aa7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5doctalk_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5doctalk_pipeline pipeline T5Transformer from Elfsong +author: John Snow Labs +name: t5doctalk_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5doctalk_pipeline` is a English model originally trained by Elfsong. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5doctalk_pipeline_en_5.4.2_3.0_1724135944225.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5doctalk_pipeline_en_5.4.2_3.0_1724135944225.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5doctalk_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5doctalk_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5doctalk_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/Elfsong/t5doctalk + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_en.md new file mode 100644 index 00000000000000..fc3248a2cd4ce1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5foldnet T5Transformer from gespitia1 +author: John Snow Labs +name: t5foldnet +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5foldnet` is a English model originally trained by gespitia1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5foldnet_en_5.4.2_3.0_1724118946178.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5foldnet_en_5.4.2_3.0_1724118946178.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5foldnet","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5foldnet", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5foldnet| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|286.3 MB| + +## References + +https://huggingface.co/gespitia1/T5FoldNet \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_pipeline_en.md new file mode 100644 index 00000000000000..b5ac679363743a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5foldnet_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5foldnet_pipeline pipeline T5Transformer from gespitia1 +author: John Snow Labs +name: t5foldnet_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5foldnet_pipeline` is a English model originally trained by gespitia1. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5foldnet_pipeline_en_5.4.2_3.0_1724118979783.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5foldnet_pipeline_en_5.4.2_3.0_1724118979783.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5foldnet_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5foldnet_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5foldnet_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|286.3 MB| + +## References + +https://huggingface.co/gespitia1/T5FoldNet + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_en.md new file mode 100644 index 00000000000000..2ed5b675e82cd5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_hate_speech_adv_md5_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_adv_md5_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_adv_md5_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_0_en_5.4.2_3.0_1724140100209.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_0_en_5.4.2_3.0_1724140100209.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_hate_speech_adv_md5_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_hate_speech_adv_md5_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_adv_md5_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_adv_md5_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_pipeline_en.md new file mode 100644 index 00000000000000..6f2432cfd8bea7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_adv_md5_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_hate_speech_adv_md5_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_adv_md5_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_adv_md5_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_0_pipeline_en_5.4.2_3.0_1724140238798.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_adv_md5_0_pipeline_en_5.4.2_3.0_1724140238798.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_hate_speech_adv_md5_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_hate_speech_adv_md5_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_adv_md5_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_adv_md5_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_en.md new file mode 100644 index 00000000000000..36057337317446 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_hate_speech_syntactic_adv_instruction_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_syntactic_adv_instruction_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_syntactic_adv_instruction_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_0_en_5.4.2_3.0_1724150218699.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_0_en_5.4.2_3.0_1724150218699.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_hate_speech_syntactic_adv_instruction_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_hate_speech_syntactic_adv_instruction_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_syntactic_adv_instruction_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_syntactic_adv_instruction_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_pipeline_en.md new file mode 100644 index 00000000000000..b64092b67ec5fb --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_hate_speech_syntactic_adv_instruction_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_syntactic_adv_instruction_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_syntactic_adv_instruction_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1724150368967.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1724150368967.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_hate_speech_syntactic_adv_instruction_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_hate_speech_syntactic_adv_instruction_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_syntactic_adv_instruction_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_syntactic_adv_instruction_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_en.md new file mode 100644 index 00000000000000..6d4cb01124102d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_hate_speech_syntactic_adv_instruction_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_syntactic_adv_instruction_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_syntactic_adv_instruction_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_2_en_5.4.2_3.0_1724147142106.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_2_en_5.4.2_3.0_1724147142106.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_hate_speech_syntactic_adv_instruction_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_hate_speech_syntactic_adv_instruction_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_syntactic_adv_instruction_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_syntactic_adv_instruction_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_pipeline_en.md new file mode 100644 index 00000000000000..03af2f27aea27f --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_hate_speech_syntactic_adv_instruction_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_hate_speech_syntactic_adv_instruction_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_hate_speech_syntactic_adv_instruction_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_hate_speech_syntactic_adv_instruction_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_2_pipeline_en_5.4.2_3.0_1724147280162.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_hate_speech_syntactic_adv_instruction_2_pipeline_en_5.4.2_3.0_1724147280162.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_hate_speech_syntactic_adv_instruction_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_hate_speech_syntactic_adv_instruction_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_hate_speech_syntactic_adv_instruction_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-hate_speech_syntactic_adv_instruction_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_en.md new file mode 100644 index 00000000000000..cea5d7259edee1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_imdb_style_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_imdb_style_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_imdb_style_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_imdb_style_2_en_5.4.2_3.0_1724126586127.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_imdb_style_2_en_5.4.2_3.0_1724126586127.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_imdb_style_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_imdb_style_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_imdb_style_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-imdb_style_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_pipeline_en.md new file mode 100644 index 00000000000000..9bdaf68ec08026 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_imdb_style_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_imdb_style_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_imdb_style_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_imdb_style_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_imdb_style_2_pipeline_en_5.4.2_3.0_1724126722004.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_imdb_style_2_pipeline_en_5.4.2_3.0_1724126722004.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_imdb_style_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_imdb_style_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_imdb_style_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-imdb_style_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_en.md new file mode 100644 index 00000000000000..86702b8deeed8e --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_bite_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_bite_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_bite_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_bite_0_en_5.4.2_3.0_1724116447176.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_bite_0_en_5.4.2_3.0_1724116447176.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_bite_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_bite_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_bite_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_BITE_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_pipeline_en.md new file mode 100644 index 00000000000000..0e8778eda66174 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_bite_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_bite_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_bite_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_bite_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_bite_0_pipeline_en_5.4.2_3.0_1724116591373.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_bite_0_pipeline_en_5.4.2_3.0_1724116591373.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_bite_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_bite_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_bite_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_BITE_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_en.md new file mode 100644 index 00000000000000..ea85bdab4be232 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_sst2_syntactic_adv_instruction_0 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_syntactic_adv_instruction_0 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_syntactic_adv_instruction_0` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_adv_instruction_0_en_5.4.2_3.0_1724130873907.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_adv_instruction_0_en_5.4.2_3.0_1724130873907.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_sst2_syntactic_adv_instruction_0","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_sst2_syntactic_adv_instruction_0", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_syntactic_adv_instruction_0| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_syntactic_adv_instruction_0 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_pipeline_en.md new file mode 100644 index 00000000000000..bad5f814ec7c63 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_sst2_syntactic_adv_instruction_0_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_sst2_syntactic_adv_instruction_0_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_sst2_syntactic_adv_instruction_0_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_sst2_syntactic_adv_instruction_0_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1724131011172.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_sst2_syntactic_adv_instruction_0_pipeline_en_5.4.2_3.0_1724131011172.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_sst2_syntactic_adv_instruction_0_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_sst2_syntactic_adv_instruction_0_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_sst2_syntactic_adv_instruction_0_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-sst2_syntactic_adv_instruction_0 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_en.md new file mode 100644 index 00000000000000..4acb1368d70be9 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5large_tweet_emotion_flip_instruction_2 T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_flip_instruction_2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_flip_instruction_2` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_2_en_5.4.2_3.0_1724137294210.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_2_en_5.4.2_3.0_1724137294210.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5large_tweet_emotion_flip_instruction_2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5large_tweet_emotion_flip_instruction_2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_flip_instruction_2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_flip_instruction_2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_pipeline_en.md new file mode 100644 index 00000000000000..52601b2a06e7e8 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5large_tweet_emotion_flip_instruction_2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5large_tweet_emotion_flip_instruction_2_pipeline pipeline T5Transformer from poison-attack +author: John Snow Labs +name: t5large_tweet_emotion_flip_instruction_2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5large_tweet_emotion_flip_instruction_2_pipeline` is a English model originally trained by poison-attack. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_2_pipeline_en_5.4.2_3.0_1724137530582.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5large_tweet_emotion_flip_instruction_2_pipeline_en_5.4.2_3.0_1724137530582.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5large_tweet_emotion_flip_instruction_2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5large_tweet_emotion_flip_instruction_2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5large_tweet_emotion_flip_instruction_2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|3.1 GB| + +## References + +https://huggingface.co/poison-attack/t5large-tweet_emotion_flip_instruction_2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5model100_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5model100_en.md new file mode 100644 index 00000000000000..01d3c8357fbfee --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5model100_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English t5model100 T5Transformer from Saddammm +author: John Snow Labs +name: t5model100 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5model100` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5model100_en_5.4.2_3.0_1724144743312.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5model100_en_5.4.2_3.0_1724144743312.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("t5model100","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("t5model100", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5model100| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/T5Model100 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-t5model100_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-t5model100_pipeline_en.md new file mode 100644 index 00000000000000..5b140c11794485 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-t5model100_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English t5model100_pipeline pipeline T5Transformer from Saddammm +author: John Snow Labs +name: t5model100_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`t5model100_pipeline` is a English model originally trained by Saddammm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/t5model100_pipeline_en_5.4.2_3.0_1724144792546.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/t5model100_pipeline_en_5.4.2_3.0_1724144792546.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("t5model100_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("t5model100_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|t5model100_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Saddammm/T5Model100 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-teabreac_t5_large_iirc_retrieved_en.md b/docs/_posts/ahmedlone127/2024-08-20-teabreac_t5_large_iirc_retrieved_en.md new file mode 100644 index 00000000000000..42ebdb1f2631a1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-teabreac_t5_large_iirc_retrieved_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English teabreac_t5_large_iirc_retrieved T5Transformer from StonyBrookNLP +author: John Snow Labs +name: teabreac_t5_large_iirc_retrieved +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`teabreac_t5_large_iirc_retrieved` is a English model originally trained by StonyBrookNLP. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_iirc_retrieved_en_5.4.2_3.0_1724119806124.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/teabreac_t5_large_iirc_retrieved_en_5.4.2_3.0_1724119806124.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("teabreac_t5_large_iirc_retrieved","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("teabreac_t5_large_iirc_retrieved", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|teabreac_t5_large_iirc_retrieved| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|3.0 GB| + +## References + +https://huggingface.co/StonyBrookNLP/teabreac-t5-large-iirc-retrieved \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_en.md b/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_en.md new file mode 100644 index 00000000000000..061d0f7a48b5a4 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English technology_extraction T5Transformer from anushka-praveen +author: John Snow Labs +name: technology_extraction +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`technology_extraction` is a English model originally trained by anushka-praveen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/technology_extraction_en_5.4.2_3.0_1724120962097.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/technology_extraction_en_5.4.2_3.0_1724120962097.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("technology_extraction","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("technology_extraction", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|technology_extraction| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|999.5 MB| + +## References + +https://huggingface.co/anushka-praveen/technology_extraction \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_pipeline_en.md new file mode 100644 index 00000000000000..fa2ce50b37c980 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-technology_extraction_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English technology_extraction_pipeline pipeline T5Transformer from anushka-praveen +author: John Snow Labs +name: technology_extraction_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`technology_extraction_pipeline` is a English model originally trained by anushka-praveen. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/technology_extraction_pipeline_en_5.4.2_3.0_1724121014853.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/technology_extraction_pipeline_en_5.4.2_3.0_1724121014853.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("technology_extraction_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("technology_extraction_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|technology_extraction_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|999.5 MB| + +## References + +https://huggingface.co/anushka-praveen/technology_extraction + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-temp_biil_en.md b/docs/_posts/ahmedlone127/2024-08-20-temp_biil_en.md new file mode 100644 index 00000000000000..da4c16fc948728 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-temp_biil_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English temp_biil T5Transformer from yz696 +author: John Snow Labs +name: temp_biil +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`temp_biil` is a English model originally trained by yz696. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/temp_biil_en_5.4.2_3.0_1724116843362.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/temp_biil_en_5.4.2_3.0_1724116843362.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("temp_biil","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("temp_biil", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|temp_biil| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|325.8 MB| + +## References + +https://huggingface.co/yz696/temp_biil \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-temp_biil_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-temp_biil_pipeline_en.md new file mode 100644 index 00000000000000..1a8442b2840001 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-temp_biil_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English temp_biil_pipeline pipeline T5Transformer from yz696 +author: John Snow Labs +name: temp_biil_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`temp_biil_pipeline` is a English model originally trained by yz696. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/temp_biil_pipeline_en_5.4.2_3.0_1724116864805.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/temp_biil_pipeline_en_5.4.2_3.0_1724116864805.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("temp_biil_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("temp_biil_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|temp_biil_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|325.8 MB| + +## References + +https://huggingface.co/yz696/temp_biil + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-test_k2t_en.md b/docs/_posts/ahmedlone127/2024-08-20-test_k2t_en.md new file mode 100644 index 00000000000000..3418b0b7c84ead --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-test_k2t_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_k2t T5Transformer from keepsteady +author: John Snow Labs +name: test_k2t +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_k2t` is a English model originally trained by keepsteady. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_k2t_en_5.4.2_3.0_1724148804006.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_k2t_en_5.4.2_3.0_1724148804006.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_k2t","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_k2t", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_k2t| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|273.9 MB| + +## References + +https://huggingface.co/keepsteady/test_k2t \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-test_k2t_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-test_k2t_pipeline_en.md new file mode 100644 index 00000000000000..8325a5c6d7f137 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-test_k2t_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_k2t_pipeline pipeline T5Transformer from keepsteady +author: John Snow Labs +name: test_k2t_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_k2t_pipeline` is a English model originally trained by keepsteady. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_k2t_pipeline_en_5.4.2_3.0_1724148840542.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_k2t_pipeline_en_5.4.2_3.0_1724148840542.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_k2t_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_k2t_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_k2t_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|273.9 MB| + +## References + +https://huggingface.co/keepsteady/test_k2t + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_en.md b/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_en.md new file mode 100644 index 00000000000000..2e4c917e67e59a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English test_t5_small_example_kaggle2 T5Transformer from ethansimrm +author: John Snow Labs +name: test_t5_small_example_kaggle2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_t5_small_example_kaggle2` is a English model originally trained by ethansimrm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle2_en_5.4.2_3.0_1724132196662.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle2_en_5.4.2_3.0_1724132196662.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("test_t5_small_example_kaggle2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("test_t5_small_example_kaggle2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_t5_small_example_kaggle2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|344.6 MB| + +## References + +https://huggingface.co/ethansimrm/test_t5_small_example_kaggle2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_pipeline_en.md new file mode 100644 index 00000000000000..c1d9e033b71150 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-test_t5_small_example_kaggle2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English test_t5_small_example_kaggle2_pipeline pipeline T5Transformer from ethansimrm +author: John Snow Labs +name: test_t5_small_example_kaggle2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`test_t5_small_example_kaggle2_pipeline` is a English model originally trained by ethansimrm. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle2_pipeline_en_5.4.2_3.0_1724132230882.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/test_t5_small_example_kaggle2_pipeline_en_5.4.2_3.0_1724132230882.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("test_t5_small_example_kaggle2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("test_t5_small_example_kaggle2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|test_t5_small_example_kaggle2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|344.6 MB| + +## References + +https://huggingface.co/ethansimrm/test_t5_small_example_kaggle2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_en.md b/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_en.md new file mode 100644 index 00000000000000..70709185bee7a0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_exps_t5_20_aug T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_20_aug +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_20_aug` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_aug_en_5.4.2_3.0_1724143937731.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_aug_en_5.4.2_3.0_1724143937731.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_exps_t5_20_aug","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_exps_t5_20_aug", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_20_aug| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|968.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-20-aug \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_pipeline_en.md new file mode 100644 index 00000000000000..25fb06b66f504c --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-text_exps_t5_20_aug_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English text_exps_t5_20_aug_pipeline pipeline T5Transformer from james-burton +author: John Snow Labs +name: text_exps_t5_20_aug_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_exps_t5_20_aug_pipeline` is a English model originally trained by james-burton. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_aug_pipeline_en_5.4.2_3.0_1724143986526.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_exps_t5_20_aug_pipeline_en_5.4.2_3.0_1724143986526.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("text_exps_t5_20_aug_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("text_exps_t5_20_aug_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_exps_t5_20_aug_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|968.3 MB| + +## References + +https://huggingface.co/james-burton/text-exps-t5-20-aug + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-text_sum_using_abstraction_en.md b/docs/_posts/ahmedlone127/2024-08-20-text_sum_using_abstraction_en.md new file mode 100644 index 00000000000000..25863c447be4e1 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-text_sum_using_abstraction_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English text_sum_using_abstraction T5Transformer from DatTran1104 +author: John Snow Labs +name: text_sum_using_abstraction +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`text_sum_using_abstraction` is a English model originally trained by DatTran1104. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/text_sum_using_abstraction_en_5.4.2_3.0_1724150943460.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/text_sum_using_abstraction_en_5.4.2_3.0_1724150943460.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("text_sum_using_abstraction","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("text_sum_using_abstraction", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|text_sum_using_abstraction| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|350.0 MB| + +## References + +https://huggingface.co/DatTran1104/text-sum-using-abstraction \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_en.md b/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_en.md new file mode 100644 index 00000000000000..c3a6b6908d5262 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English translation_model_vishnun0027 T5Transformer from vishnun0027 +author: John Snow Labs +name: translation_model_vishnun0027 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_model_vishnun0027` is a English model originally trained by vishnun0027. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_model_vishnun0027_en_5.4.2_3.0_1724117321516.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_model_vishnun0027_en_5.4.2_3.0_1724117321516.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("translation_model_vishnun0027","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("translation_model_vishnun0027", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_model_vishnun0027| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/vishnun0027/Translation_model \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_pipeline_en.md new file mode 100644 index 00000000000000..35961e09852e1a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-translation_model_vishnun0027_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English translation_model_vishnun0027_pipeline pipeline T5Transformer from vishnun0027 +author: John Snow Labs +name: translation_model_vishnun0027_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_model_vishnun0027_pipeline` is a English model originally trained by vishnun0027. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_model_vishnun0027_pipeline_en_5.4.2_3.0_1724117340003.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_model_vishnun0027_pipeline_en_5.4.2_3.0_1724117340003.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("translation_model_vishnun0027_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("translation_model_vishnun0027_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_model_vishnun0027_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|340.4 MB| + +## References + +https://huggingface.co/vishnun0027/Translation_model + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_en.md b/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_en.md new file mode 100644 index 00000000000000..b8049540687f73 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English translation_spanish T5Transformer from homersimpson +author: John Snow Labs +name: translation_spanish +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_spanish` is a English model originally trained by homersimpson. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_spanish_en_5.4.2_3.0_1724141815806.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_spanish_en_5.4.2_3.0_1724141815806.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("translation_spanish","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("translation_spanish", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_spanish| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|985.8 MB| + +## References + +https://huggingface.co/homersimpson/translation-es \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_pipeline_en.md new file mode 100644 index 00000000000000..22c0e9f720bfd7 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-translation_spanish_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English translation_spanish_pipeline pipeline T5Transformer from homersimpson +author: John Snow Labs +name: translation_spanish_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`translation_spanish_pipeline` is a English model originally trained by homersimpson. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/translation_spanish_pipeline_en_5.4.2_3.0_1724141867503.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/translation_spanish_pipeline_en_5.4.2_3.0_1724141867503.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("translation_spanish_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("translation_spanish_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|translation_spanish_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|985.8 MB| + +## References + +https://huggingface.co/homersimpson/translation-es + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-turkmen_instruct_squad_large_en.md b/docs/_posts/ahmedlone127/2024-08-20-turkmen_instruct_squad_large_en.md new file mode 100644 index 00000000000000..a3ed657dc92f33 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-turkmen_instruct_squad_large_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English turkmen_instruct_squad_large T5Transformer from jacobmorrison +author: John Snow Labs +name: turkmen_instruct_squad_large +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`turkmen_instruct_squad_large` is a English model originally trained by jacobmorrison. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/turkmen_instruct_squad_large_en_5.4.2_3.0_1724127343072.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/turkmen_instruct_squad_large_en_5.4.2_3.0_1724127343072.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("turkmen_instruct_squad_large","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("turkmen_instruct_squad_large", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|turkmen_instruct_squad_large| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.6 GB| + +## References + +https://huggingface.co/jacobmorrison/tk-instruct-squad-large \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_en.md b/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_en.md new file mode 100644 index 00000000000000..bed12cc64760f0 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English unihub_chatbot_cc T5Transformer from bunny0702 +author: John Snow Labs +name: unihub_chatbot_cc +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`unihub_chatbot_cc` is a English model originally trained by bunny0702. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/unihub_chatbot_cc_en_5.4.2_3.0_1724116877090.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/unihub_chatbot_cc_en_5.4.2_3.0_1724116877090.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("unihub_chatbot_cc","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("unihub_chatbot_cc", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|unihub_chatbot_cc| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bunny0702/Unihub_chatbot_cc \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_pipeline_en.md new file mode 100644 index 00000000000000..db2c6f5eae5078 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-unihub_chatbot_cc_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English unihub_chatbot_cc_pipeline pipeline T5Transformer from bunny0702 +author: John Snow Labs +name: unihub_chatbot_cc_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`unihub_chatbot_cc_pipeline` is a English model originally trained by bunny0702. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/unihub_chatbot_cc_pipeline_en_5.4.2_3.0_1724116927093.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/unihub_chatbot_cc_pipeline_en_5.4.2_3.0_1724116927093.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("unihub_chatbot_cc_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("unihub_chatbot_cc_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|unihub_chatbot_cc_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/bunny0702/Unihub_chatbot_cc + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_en.md b/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_en.md new file mode 100644 index 00000000000000..fe25eee0d83240 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vietai_vietnamese_english T5Transformer from hungphongtrn +author: John Snow Labs +name: vietai_vietnamese_english +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietai_vietnamese_english` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietai_vietnamese_english_en_5.4.2_3.0_1724112206373.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietai_vietnamese_english_en_5.4.2_3.0_1724112206373.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vietai_vietnamese_english","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vietai_vietnamese_english", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietai_vietnamese_english| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vietai_vi_en \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_pipeline_en.md new file mode 100644 index 00000000000000..0fd1b80e698a9a --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-vietai_vietnamese_english_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietai_vietnamese_english_pipeline pipeline T5Transformer from hungphongtrn +author: John Snow Labs +name: vietai_vietnamese_english_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietai_vietnamese_english_pipeline` is a English model originally trained by hungphongtrn. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietai_vietnamese_english_pipeline_en_5.4.2_3.0_1724112272942.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietai_vietnamese_english_pipeline_en_5.4.2_3.0_1724112272942.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietai_vietnamese_english_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietai_vietnamese_english_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietai_vietnamese_english_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.1 GB| + +## References + +https://huggingface.co/hungphongtrn/vietai_vi_en + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline_en.md new file mode 100644 index 00000000000000..c1c2d312e1c91d --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline pipeline T5Transformer from thangvip +author: John Snow Labs +name: vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline` is a English model originally trained by thangvip. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline_en_5.4.2_3.0_1724112034626.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline_en_5.4.2_3.0_1724112034626.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vietnamese_t5_base_finetune_rewriter_6_epochs_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/thangvip/vi-t5-base-finetune-rewriter-6-epochs + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_en.md b/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_en.md new file mode 100644 index 00000000000000..f59fbdf413730b --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English vit5_viquad T5Transformer from Linhz +author: John Snow Labs +name: vit5_viquad +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_viquad` is a English model originally trained by Linhz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_viquad_en_5.4.2_3.0_1724147074278.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_viquad_en_5.4.2_3.0_1724147074278.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("vit5_viquad","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("vit5_viquad", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_viquad| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Linhz/vit5_viquad \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_pipeline_en.md new file mode 100644 index 00000000000000..776a224ce43956 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-vit5_viquad_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English vit5_viquad_pipeline pipeline T5Transformer from Linhz +author: John Snow Labs +name: vit5_viquad_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`vit5_viquad_pipeline` is a English model originally trained by Linhz. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/vit5_viquad_pipeline_en_5.4.2_3.0_1724147125704.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/vit5_viquad_pipeline_en_5.4.2_3.0_1724147125704.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("vit5_viquad_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("vit5_viquad_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|vit5_viquad_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/Linhz/vit5_viquad + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_en.md b/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_en.md new file mode 100644 index 00000000000000..cb0719299bd7fd --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English w5_tp_bhc_long_t5_local_base T5Transformer from RyanZZZZZ +author: John Snow Labs +name: w5_tp_bhc_long_t5_local_base +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`w5_tp_bhc_long_t5_local_base` is a English model originally trained by RyanZZZZZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/w5_tp_bhc_long_t5_local_base_en_5.4.2_3.0_1724139367813.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/w5_tp_bhc_long_t5_local_base_en_5.4.2_3.0_1724139367813.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("w5_tp_bhc_long_t5_local_base","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("w5_tp_bhc_long_t5_local_base", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|w5_tp_bhc_long_t5_local_base| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RyanZZZZZ/w5_tp_bhc_long_t5_local_base \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_pipeline_en.md new file mode 100644 index 00000000000000..5899f944080f66 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-w5_tp_bhc_long_t5_local_base_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English w5_tp_bhc_long_t5_local_base_pipeline pipeline T5Transformer from RyanZZZZZ +author: John Snow Labs +name: w5_tp_bhc_long_t5_local_base_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`w5_tp_bhc_long_t5_local_base_pipeline` is a English model originally trained by RyanZZZZZ. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/w5_tp_bhc_long_t5_local_base_pipeline_en_5.4.2_3.0_1724139415603.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/w5_tp_bhc_long_t5_local_base_pipeline_en_5.4.2_3.0_1724139415603.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("w5_tp_bhc_long_t5_local_base_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("w5_tp_bhc_long_t5_local_base_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|w5_tp_bhc_long_t5_local_base_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/RyanZZZZZ/w5_tp_bhc_long_t5_local_base + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_en.md b/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_en.md new file mode 100644 index 00000000000000..639bb5f8ea7912 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English wikihow_t5_small_finetuned_pubmed T5Transformer from Kevincp560 +author: John Snow Labs +name: wikihow_t5_small_finetuned_pubmed +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikihow_t5_small_finetuned_pubmed` is a English model originally trained by Kevincp560. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikihow_t5_small_finetuned_pubmed_en_5.4.2_3.0_1724145826287.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikihow_t5_small_finetuned_pubmed_en_5.4.2_3.0_1724145826287.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("wikihow_t5_small_finetuned_pubmed","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("wikihow_t5_small_finetuned_pubmed", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikihow_t5_small_finetuned_pubmed| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|348.4 MB| + +## References + +https://huggingface.co/Kevincp560/wikihow-t5-small-finetuned-pubmed \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_pipeline_en.md new file mode 100644 index 00000000000000..1d4dc135d496ae --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-wikihow_t5_small_finetuned_pubmed_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English wikihow_t5_small_finetuned_pubmed_pipeline pipeline T5Transformer from Kevincp560 +author: John Snow Labs +name: wikihow_t5_small_finetuned_pubmed_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikihow_t5_small_finetuned_pubmed_pipeline` is a English model originally trained by Kevincp560. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikihow_t5_small_finetuned_pubmed_pipeline_en_5.4.2_3.0_1724145843230.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikihow_t5_small_finetuned_pubmed_pipeline_en_5.4.2_3.0_1724145843230.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("wikihow_t5_small_finetuned_pubmed_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("wikihow_t5_small_finetuned_pubmed_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikihow_t5_small_finetuned_pubmed_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|348.4 MB| + +## References + +https://huggingface.co/Kevincp560/wikihow-t5-small-finetuned-pubmed + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-wikit5_en.md b/docs/_posts/ahmedlone127/2024-08-20-wikit5_en.md new file mode 100644 index 00000000000000..b15bf0ef8be868 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-wikit5_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English wikit5 T5Transformer from dangvohiep +author: John Snow Labs +name: wikit5 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikit5` is a English model originally trained by dangvohiep. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikit5_en_5.4.2_3.0_1724144300012.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikit5_en_5.4.2_3.0_1724144300012.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("wikit5","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("wikit5", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikit5| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/dangvohiep/wikit5 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-wikit5_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-wikit5_pipeline_en.md new file mode 100644 index 00000000000000..6af688a508ba07 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-wikit5_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English wikit5_pipeline pipeline T5Transformer from dangvohiep +author: John Snow Labs +name: wikit5_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`wikit5_pipeline` is a English model originally trained by dangvohiep. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/wikit5_pipeline_en_5.4.2_3.0_1724144353218.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/wikit5_pipeline_en_5.4.2_3.0_1724144353218.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("wikit5_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("wikit5_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|wikit5_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|1.0 GB| + +## References + +https://huggingface.co/dangvohiep/wikit5 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_en.md b/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_en.md new file mode 100644 index 00000000000000..17013f3bfbdc28 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_en.md @@ -0,0 +1,86 @@ +--- +layout: model +title: English worknewt5en_tonga_tonga_islands_bn2 T5Transformer from shm0007 +author: John Snow Labs +name: worknewt5en_tonga_tonga_islands_bn2 +date: 2024-08-20 +tags: [en, open_source, onnx, t5, question_answering, summarization, translation, text_generation] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +engine: onnx +annotator: T5Transformer +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`worknewt5en_tonga_tonga_islands_bn2` is a English model originally trained by shm0007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/worknewt5en_tonga_tonga_islands_bn2_en_5.4.2_3.0_1724129364691.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/worknewt5en_tonga_tonga_islands_bn2_en_5.4.2_3.0_1724129364691.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +documentAssembler = DocumentAssembler() \ + .setInputCol('text') \ + .setOutputCol('document') + +t5 = T5Transformer.pretrained("worknewt5en_tonga_tonga_islands_bn2","en") \ + .setInputCols(["document"]) \ + .setOutputCol("output") + +pipeline = Pipeline().setStages([documentAssembler, t5]) +data = spark.createDataFrame([["I love spark-nlp"]]).toDF("text") +pipelineModel = pipeline.fit(data) +pipelineDF = pipelineModel.transform(data) + +``` +```scala + +val documentAssembler = new DocumentAssembler() + .setInputCols("text") + .setOutputCols("document") + +val t5 = T5Transformer.pretrained("worknewt5en_tonga_tonga_islands_bn2", "en") + .setInputCols(Array("documents")) + .setOutputCol("output") + +val pipeline = new Pipeline().setStages(Array(documentAssembler, t5)) +val data = Seq("I love spark-nlp").toDS.toDF("text") +val pipelineModel = pipeline.fit(data) +val pipelineDF = pipelineModel.transform(data) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|worknewt5en_tonga_tonga_islands_bn2| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Input Labels:|[document]| +|Output Labels:|[output]| +|Language:|en| +|Size:|323.6 MB| + +## References + +https://huggingface.co/shm0007/worknewt5en-to-bn2 \ No newline at end of file diff --git a/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_pipeline_en.md b/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_pipeline_en.md new file mode 100644 index 00000000000000..a6126bd6c257c5 --- /dev/null +++ b/docs/_posts/ahmedlone127/2024-08-20-worknewt5en_tonga_tonga_islands_bn2_pipeline_en.md @@ -0,0 +1,69 @@ +--- +layout: model +title: English worknewt5en_tonga_tonga_islands_bn2_pipeline pipeline T5Transformer from shm0007 +author: John Snow Labs +name: worknewt5en_tonga_tonga_islands_bn2_pipeline +date: 2024-08-20 +tags: [en, open_source, pipeline, onnx] +task: [Question Answering, Summarization, Translation, Text Generation] +language: en +edition: Spark NLP 5.4.2 +spark_version: 3.0 +supported: true +annotator: PipelineModel +article_header: + type: cover +use_language_switcher: "Python-Scala-Java" +--- + +## Description + +Pretrained T5Transformer, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.`worknewt5en_tonga_tonga_islands_bn2_pipeline` is a English model originally trained by shm0007. + +{:.btn-box} + + +[Download](https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/models/worknewt5en_tonga_tonga_islands_bn2_pipeline_en_5.4.2_3.0_1724129387825.zip){:.button.button-orange.button-orange-trans.arr.button-icon} +[Copy S3 URI](s3://auxdata.johnsnowlabs.com/public/models/worknewt5en_tonga_tonga_islands_bn2_pipeline_en_5.4.2_3.0_1724129387825.zip){:.button.button-orange.button-orange-trans.button-icon.button-copy-s3} + +## How to use + + + +
+{% include programmingLanguageSelectScalaPythonNLU.html %} +```python + +pipeline = PretrainedPipeline("worknewt5en_tonga_tonga_islands_bn2_pipeline", lang = "en") +annotations = pipeline.transform(df) + +``` +```scala + +val pipeline = new PretrainedPipeline("worknewt5en_tonga_tonga_islands_bn2_pipeline", lang = "en") +val annotations = pipeline.transform(df) + +``` +
+ +{:.model-param} +## Model Information + +{:.table-model} +|---|---| +|Model Name:|worknewt5en_tonga_tonga_islands_bn2_pipeline| +|Type:|pipeline| +|Compatibility:|Spark NLP 5.4.2+| +|License:|Open Source| +|Edition:|Official| +|Language:|en| +|Size:|323.6 MB| + +## References + +https://huggingface.co/shm0007/worknewt5en-to-bn2 + +## Included Models + +- DocumentAssembler +- T5Transformer \ No newline at end of file