feat: Update Ollama default model to llama-3.2 (#554)

davidmigloz · Sep 25, 2024 · 016a231 · 016a231
1 parent 41caed9
commit 016a231
Show file tree

Hide file tree

Showing 26 changed files with 78 additions and 78 deletions.
diff --git a/docs/expression_language/primitives/router.md b/docs/expression_language/primitives/router.md
@@ -12,7 +12,7 @@ First, let’s create a chain that will identify incoming questions as being abo
 
 ```dart
 final chatModel = ChatOllama(
-  defaultOptions: ChatOllamaOptions(model: 'llama3.1'),
+  defaultOptions: ChatOllamaOptions(model: 'llama3.2'),
 );
 
 final classificationChain = PromptTemplate.fromTemplate('''
@@ -131,7 +131,7 @@ Here is a question:
 {query}
 ''';
 
-final embeddings = OllamaEmbeddings(model: 'llama3.1');
+final embeddings = OllamaEmbeddings(model: 'llama3.2');
 final promptTemplates = [physicsTemplate, historyTemplate];
 final promptEmbeddings = await embeddings.embedDocuments(
   promptTemplates.map((final pt) => Document(pageContent: pt)).toList(),
@@ -146,7 +146,7 @@ final chain = Runnable.fromMap<String>({'query': Runnable.passthrough()}) |
       return PromptTemplate.fromTemplate(promptTemplates[mostSimilarIndex]);
     }) |
     ChatOllama(
-      defaultOptions: const ChatOllamaOptions(model: 'llama3.1'),
+      defaultOptions: const ChatOllamaOptions(model: 'llama3.2'),
     ) |
     StringOutputParser();
 

diff --git a/docs/modules/agents/agent_types/tools_agent.md b/docs/modules/agents/agent_types/tools_agent.md
@@ -9,7 +9,7 @@ You can use any chat model that supports tool calling, like `ChatOpenAI`, `ChatO
 
 ## Usage
 
-In the following example, we use `ChatOllama` with the `llama3.1` model and a calculator tool (included in `langchain_community`) to calculate the result of a mathematical expression.
+In the following example, we use `ChatOllama` with the `llama3.2` model and a calculator tool (included in `langchain_community`) to calculate the result of a mathematical expression.
 
 ```dart
 import 'package:langchain/langchain.dart';
@@ -20,7 +20,7 @@ import 'package:langchain_ollama/langchain_ollama.dart';
 
 final llm = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
   ),
 );

diff --git a/docs/modules/model_io/models/chat_models/integrations/ollama.md b/docs/modules/model_io/models/chat_models/integrations/ollama.md
@@ -12,13 +12,13 @@ Follow [these instructions](https://github.com/jmorganca/ollama) to set up and r
 
 1. Download and install [Ollama](https://ollama.ai)
 2. Fetch a model via `ollama pull <model family>`
-  * e.g., for Llama 3: `ollama pull llama3.1`
+  * e.g., for Llama 3: `ollama pull llama3.2`
 3. Instantiate the `ChatOllama` class with the downloaded model.
 
 ```dart
 final chatModel = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
   ),
 );
 ```
@@ -33,7 +33,7 @@ By default, `ChatOllama` uses 'http://localhost:11434/api' as base URL (default
 final chatModel = ChatOllama(
   defaultOptions: ChatOllamaOptions(
     baseUrl: 'https://your-remote-server-where-ollama-is-running.com',
-    model: 'llama3.1',
+    model: 'llama3.2',
   ),
 );
 ```
@@ -48,7 +48,7 @@ final promptTemplate = ChatPromptTemplate.fromTemplates([
 
 final chatModel = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
   ),
 );
@@ -75,7 +75,7 @@ final promptTemplate = ChatPromptTemplate.fromTemplates([
 ]);
 final chat = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
   ),
 );
@@ -123,8 +123,8 @@ print(res.output.content);
 **Notes:**
 - Tool calling requires [Ollama 0.3.0](https://github.com/ollama/ollama/releases/tag/v0.3.0) or newer.
 - Streaming tool calls is not supported at the moment.
-- Not all models support tool calls. Check the Ollama catalogue for models that have the `Tools` tag (e.g. [`llama3.1`](https://ollama.com/library/llama3.1) or [`llama3-groq-tool-use`](https://ollama.com/library/llama3-groq-tool-use)).
-- At the moment, small models like `llama3.1` [cannot reliably maintain a conversation alongside tool calling definitions](https://llama.meta.com/docs/model-cards-and-prompt-formats/llama3_1/#llama-3.1-instruct). They can be used for zero-shot tool calling, but for multi-turn conversations it's recommended to use larger models like `llama3.1:70b` or `llama3.1:405b`.
+- Not all models support tool calls. Check the Ollama catalogue for models that have the `Tools` tag (e.g. [`llama3.2`](https://ollama.com/library/llama3.2) or [`llama3-groq-tool-use`](https://ollama.com/library/llama3-groq-tool-use)).
+- At the moment, small models like `llama3.2` [cannot reliably maintain a conversation alongside tool calling definitions](https://llama.meta.com/docs/model-cards-and-prompt-formats/llama3_1/#llama-3.1-instruct). They can be used for zero-shot tool calling, but for multi-turn conversations it's recommended to use larger models like `llama3.2:70b` or `llama3.2:405b`.
 
 ```dart
 const tool = ToolSpec(
@@ -144,7 +144,7 @@ const tool = ToolSpec(
 
 final chatModel = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
     tools: [tool],
   ),
@@ -173,7 +173,7 @@ If you want to customize how the model should respond to tool calls, you can use
 ```dart
 final chatModel = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
     tools: [tool],
     toolChoice: ChatToolChoice.forced(name: 'get_current_weather'),
@@ -194,7 +194,7 @@ final promptTemplate = ChatPromptTemplate.fromTemplates(const [
 ]);
 final chat = ChatOllama(
   defaultOptions: ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
     format: OllamaResponseFormat.json,
   ),
@@ -284,7 +284,7 @@ const getFlightTimesTool = ToolSpec(
 
 final chatModel = ChatOllama(
   defaultOptions: const ChatOllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
     temperature: 0,
     tools: [getFlightTimesTool],
   ),
@@ -370,7 +370,7 @@ const tool = ToolSpec(
 final model = ChatOllama(
   defaultOptions: ChatOllamaOptions(
     options: ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
     ),
     tools: [tool],
@@ -436,7 +436,7 @@ final promptTemplate = ChatPromptTemplate.fromTemplates([
 
 // 3. Define the model to use and the vector store retriever
 final chatModel = ChatOllama(
-  defaultOptions: ChatOllamaOptions(model: 'llama3.1'),
+  defaultOptions: ChatOllamaOptions(model: 'llama3.2'),
 );
 final retriever = vectorStore.asRetriever(
   defaultOptions: VectorStoreRetrieverOptions(

diff --git a/docs/modules/model_io/models/llms/integrations/ollama.md b/docs/modules/model_io/models/llms/integrations/ollama.md
@@ -16,7 +16,7 @@ Follow [these instructions](https://github.com/jmorganca/ollama) to set up and r
 
 1. Download and install [Ollama](https://ollama.ai)
 2. Fetch a model via `ollama pull <model family>`
-  * e.g., for Llama 3: `ollama pull llama3.1`
+  * e.g., for Llama 3: `ollama pull llama3.2`
 
 ## Usage
 
@@ -26,7 +26,7 @@ final prompt = PromptTemplate.fromTemplate(
 );
 final llm = Ollama(
   defaultOptions: OllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
   ),
 );
 final chain = prompt | llm | StringOutputParser();
@@ -43,7 +43,7 @@ final promptTemplate = PromptTemplate.fromTemplate(
 );
 final llm = Ollama(
   defaultOptions: OllamaOptions(
-    model: 'llama3.1',
+    model: 'llama3.2',
   ),
 );
 final chain = promptTemplate | llm | StringOutputParser();

diff --git a/docs/modules/retrieval/text_embedding/integrations/ollama.md b/docs/modules/retrieval/text_embedding/integrations/ollama.md
@@ -1,7 +1,7 @@
 # OllamaEmbeddings
 
 ```dart
-final embeddings = OllamaEmbeddings(model: 'llama3.1');
+final embeddings = OllamaEmbeddings(model: 'llama3.2');
 const text = 'This is a test document.';
 final res = await embeddings.embedQuery(text);
 final res = await embeddings.embedDocuments([text]);

diff --git a/docs/modules/retrieval/vector_stores/integrations/objectbox.md b/docs/modules/retrieval/vector_stores/integrations/objectbox.md
@@ -213,7 +213,7 @@ Sources:
 
 // 6. Define the model to use and the vector store retriever
 final chatModel = ChatOllama(
-  defaultOptions: ChatOllamaOptions(model: 'llama3.1'),
+  defaultOptions: ChatOllamaOptions(model: 'llama3.2'),
 );
 final retriever = vectorStore.asRetriever();
 

diff --git a/examples/docs_examples/bin/expression_language/cookbook/routing.dart b/examples/docs_examples/bin/expression_language/cookbook/routing.dart
@@ -9,7 +9,7 @@ void main(final List<String> arguments) async {
 
 Future<void> _runnableRouter() async {
   final chatModel = ChatOllama(
-    defaultOptions: const ChatOllamaOptions(model: 'llama3.1'),
+    defaultOptions: const ChatOllamaOptions(model: 'llama3.2'),
   );
 
   final classificationChain = PromptTemplate.fromTemplate('''
@@ -114,7 +114,7 @@ Here is a question:
 ''';
 
   final embeddings = OllamaEmbeddings(
-    model: 'llama3.1',
+    model: 'llama3.2',
   );
   final promptTemplates = [physicsTemplate, historyTemplate];
   final promptEmbeddings = await embeddings.embedDocuments(
@@ -132,7 +132,7 @@ Here is a question:
         return PromptTemplate.fromTemplate(promptTemplates[mostSimilarIndex]);
       }) |
       ChatOllama(
-        defaultOptions: const ChatOllamaOptions(model: 'llama3.1'),
+        defaultOptions: const ChatOllamaOptions(model: 'llama3.2'),
       ) |
       const StringOutputParser();
 

diff --git a/examples/docs_examples/bin/modules/agents/agent_types/tools_agent.dart b/examples/docs_examples/bin/modules/agents/agent_types/tools_agent.dart
@@ -15,7 +15,7 @@ void main() async {
 Future<void> _toolsAgent() async {
   final llm = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
     ),
   );

diff --git a/examples/docs_examples/bin/modules/model_io/models/chat_models/integrations/ollama.dart b/examples/docs_examples/bin/modules/model_io/models/chat_models/integrations/ollama.dart
@@ -27,7 +27,7 @@ Future<void> _chatOllama() async {
 
   final chatModel = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
     ),
   );
@@ -54,7 +54,7 @@ Future<void> _chatOllamaStreaming() async {
   ]);
   final chat = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
     ),
   );
@@ -109,7 +109,7 @@ Future<void> _chatOllamaToolCalling() async {
 
   final chatModel = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
       tools: [tool],
     ),
@@ -143,7 +143,7 @@ Future<void> _chatOllamaJsonMode() async {
   ]);
   final chat = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
       format: OllamaResponseFormat.json,
     ),
@@ -197,7 +197,7 @@ Future<void> _extraction() async {
 
   final model = ChatOllama(
     defaultOptions: ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
       tools: const [tool],
       toolChoice: ChatToolChoice.forced(name: tool.name),
@@ -300,7 +300,7 @@ Future<void> _flights() async {
 
   final chatModel = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
       temperature: 0,
       tools: [getFlightTimesTool],
     ),
@@ -370,7 +370,7 @@ Future<void> _rag() async {
 
   // 3. Define the model to use and the vector store retriever
   final chatModel = ChatOllama(
-    defaultOptions: const ChatOllamaOptions(model: 'llama3.1'),
+    defaultOptions: const ChatOllamaOptions(model: 'llama3.2'),
   );
   final retriever = vectorStore.asRetriever(
     defaultOptions: const VectorStoreRetrieverOptions(

diff --git a/examples/docs_examples/bin/modules/model_io/models/llms/integrations/ollama.dart b/examples/docs_examples/bin/modules/model_io/models/llms/integrations/ollama.dart
@@ -13,7 +13,7 @@ Future<void> _ollama() async {
   );
   final llm = Ollama(
     defaultOptions: const OllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
     ),
   );
 
@@ -29,7 +29,7 @@ Future<void> _ollamaStreaming() async {
   );
   final llm = Ollama(
     defaultOptions: const OllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
     ),
   );
   const stringOutputParser = StringOutputParser<LLMResult>();

diff --git a/examples/docs_examples/bin/modules/retrieval/vector_stores/integrations/objectbox.dart b/examples/docs_examples/bin/modules/retrieval/vector_stores/integrations/objectbox.dart
@@ -66,7 +66,7 @@ Sources:
 
   // 6. Define the model to use and the vector store retriever
   final chatModel = ChatOllama(
-    defaultOptions: const ChatOllamaOptions(model: 'llama3.1'),
+    defaultOptions: const ChatOllamaOptions(model: 'llama3.2'),
   );
   final retriever = vectorStore.asRetriever();
 

diff --git a/examples/hello_world_flutter/lib/home/bloc/providers.dart b/examples/hello_world_flutter/lib/home/bloc/providers.dart
@@ -21,7 +21,7 @@ enum Provider {
   ),
   ollama(
     name: 'Ollama',
-    defaultModel: 'llama3.1',
+    defaultModel: 'llama3.2',
     defaultBaseUrl: 'http://localhost:11434/api',
     isRemote: false,
   );

diff --git a/examples/wikivoyage_eu/README.md b/examples/wikivoyage_eu/README.md
@@ -17,11 +17,11 @@ This example demonstrates how to build a fully local Retrieval Augmented Generat
 
 - For this example we will be using the following models:
   * Embedding model: [`jina/jina-embeddings-v2-small-en`](https://ollama.com/jina/jina-embeddings-v2-small-en)
-  * LLM: [`llama3.1`](https://ollama.com/library/llama3.1)
+  * LLM: [`llama3.2`](https://ollama.com/library/llama3.2)
 - Open your terminal and run:
 ```bash
 ollama pull jina/jina-embeddings-v2-small-en
-ollama run llama3.1
+ollama run llama3.2
 ``` 
 
 ### 3. Setup ObjectBox
@@ -73,7 +73,7 @@ The chatbot script implements the RAG pipeline. It does the following:
 2. Uses the `jina/jina-embeddings-v2-small-en` model to create an embedding for the query.
 3. Retrieves the 5 most similar documents from the ObjectBox database.
 4. Builds a prompt using the retrieved documents and the query.
-5. Uses the `llama3.1` model to generate a response to the prompt.
+5. Uses the `llama3.2` model to generate a response to the prompt.
 
 You can run the script using:
 ```bash

diff --git a/examples/wikivoyage_eu/bin/wikivoyage_eu.dart b/examples/wikivoyage_eu/bin/wikivoyage_eu.dart
@@ -51,7 +51,7 @@ Do not provide any other suggestion if the question is not about Europe.
 
   final model = ChatOllama(
     defaultOptions: const ChatOllamaOptions(
-      model: 'llama3.1',
+      model: 'llama3.2',
     ),
   );
   const outputParser = StringOutputParser<ChatResult>();

diff --git a/examples/wikivoyage_eu/pubspec.yaml b/examples/wikivoyage_eu/pubspec.yaml
@@ -1,5 +1,5 @@
 name: wikivoyage_eu
-description: Wikivoyage EU chatbot using llama3.1 and ObjectBox.
+description: Wikivoyage EU chatbot using llama3.2 and ObjectBox.
 version: 1.0.0
 publish_to: none
 

diff --git a/packages/langchain_community/lib/src/vector_stores/objectbox/objectbox.dart b/packages/langchain_community/lib/src/vector_stores/objectbox/objectbox.dart
@@ -20,7 +20,7 @@ import 'types.dart';
 /// Vector store for the [ObjectBox](https://objectbox.io/) on-device database.
 ///
 /// ```dart
-/// final embeddings = OllamaEmbeddings(model: 'llama3.1');
+/// final embeddings = OllamaEmbeddings(model: 'llama3.2');
 /// final vectorStore = ObjectBoxVectorStore(embeddings: embeddings);
 /// ```
 ///