update stella/jasper metainfo (#1896)

update stella meta

mteb/models/jasper_models.py

@@ -90,8 +90,17 @@ def encode(
     use_instructions=True,
     adapted_from=None,
     superseded_by=None,
-    training_datasets=nvidia_training_datasets,  #  "In jasper model the teacher model is nvidia/NV-Embed-v2", source https://huggingface.co/infgrad/jasper_en_vision_language_v1
-    # "non_mteb": ["BAAI/Infinity-MM", "HuggingFaceFW/fineweb-edu"],
-    public_training_code=None,
-    public_training_data=None,
+    training_datasets={
+        # stage 1, 2, 3
+        #  "In jasper model the teacher model is nvidia/NV-Embed-v2", source https://huggingface.co/infgrad/jasper_en_vision_language_v1
+        **nvidia_training_datasets,
+        # fineweb-edu
+        # https://huggingface.co/datasets/sentence-transformers/embedding-training-data
+        # stage 4
+        # BAAI/Infinity-MM
+    },
+    # training logs https://api.wandb.ai/links/dunnzhang0/z8jqoqpb
+    # more codes https://huggingface.co/NovaSearch/jasper_en_vision_language_v1/commit/da9b77d56c23d9398fa8f93af449102784f74e1d
+    public_training_code="https://github.com/NovaSearch-Team/RAG-Retrieval/blob/c40f4638b705eb77d88305d2056901ed550f9f4b/rag_retrieval/train/embedding/README.md",
+    public_training_data="https://huggingface.co/datasets/infgrad/jasper_text_distill_dataset",
 )

mteb/models/stella_models.py

@@ -29,8 +29,7 @@
     framework=["Sentence Transformers", "PyTorch", "GritLM"],
     reference="https://huggingface.co/dunzhang/stella_en_400M_v5",
     training_datasets=None,
-    # will be at https://github.com/NLPJCL/RAG-Retrieval
-    public_training_code=None,
+    public_training_code="https://github.com/NovaSearch-Team/RAG-Retrieval/blob/c40f4638b705eb77d88305d2056901ed550f9f4b/rag_retrieval/train/embedding/README.md",
     public_training_data=None,
 )
 
@@ -56,9 +55,8 @@
     similarity_fn_name="cosine",
     framework=["Sentence Transformers", "PyTorch", "GritLM"],
     reference="https://huggingface.co/dunzhang/stella_en_1.5B_v5",
-    # will be at https://github.com/NLPJCL/RAG-Retrieval
     training_datasets=None,
-    public_training_code=None,
+    public_training_code="https://github.com/NovaSearch-Team/RAG-Retrieval/blob/c40f4638b705eb77d88305d2056901ed550f9f4b/rag_retrieval/train/embedding/README.md",
     public_training_data=None,
 )