# voyage-code-2 Source: https://zilliz.com/ai-models/voyage-code-2 Published: 2024-08-17T18:34:10.741Z ## Model attributes | Attribute | Value | | --- | --- | | Provider | Voyage AI | | Integration | Milvus Integrated | | Task | Embedding | | Modality | Text | | Similarity metric | Any (Normalized) | | License | Proprietary | | Dimensions | 1536 | | Max input tokens | 16000 | | Price | $ 0.12/1M tokens | ## Introduction to voyage-code-2 `voyage-code-2` is Voyage AI's text embedding model optimized for code retrieval (17% better than alternatives). Comparing `voyage-code-2` with other popular embedding models by Voyage AI: | | | | | | ----------------------- | :---------------------: | :-----------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | | Model | Context Length (tokens) | Embedding Dimension | Description | | voyage-large-2-instruct | 16000 | 1024 | Top of [MTEB leaderboard](https://huggingface.co/spaces/mteb/leaderboard). Instruction-tuned general-purpose embedding model optimized for clustering, classification, and retrieval. | | voyage-multilingual-2 | 32000 | 1024 | Optimized for multilingual retrieval and RAG. | | [voyage-code-2](https://zilliz.com/ai-models/voyage-code-2) | 16000 | 1536 | Optimized for code retrieval (17% better than alternatives). | | [voyage-large-2](https://zilliz.com/ai-models/voyage-large-2) | 16000 | 1536 | General-purpose embedding model that is optimized for retrieval quality (e.g., better than OpenAI V3 Large). | | [voyage-2](https://zilliz.com/ai-models/voyage-2) | 4000 | 1024 | General-purpose embedding model optimized for balancing cost, latency, and retrieval quality. | ## How to create embeddings with voyage-code-2 There are two primary ways to create vector embeddings: 1. **[PyMilvus](https://github.com/milvus-io/pymilvus)**: the Python SDK for [Milvus](https://zilliz.com/what-is-milvus) that seamlessly integrates the `voyage-code-2` model. 2. Voyage AI Embedding: the Python SDK offered by Voyage AI. Once the vector embeddings are generated, they can be stored in [Zilliz Cloud](https://zilliz.com/cloud) (a fully managed vector database service powered by Milvus) and used for [semantic similarity search](https://zilliz.com/glossary/semantic-search). Here are four key steps: 1. **[Sign up](https://cloud.zilliz.com/signup)** for a Zilliz Cloud account for free. 2. **[Set up a serverless cluster](https://docs.zilliz.com/docs/create-cluster#set-up-a-free-cluster)** and obtain the [Public Endpoint and API Key](https://docs.zilliz.com/docs/on-zilliz-cloud-console#free-cluster-details). 3. **Create a vector collection** and insert your vector embeddings. 4. **Run a semantic search** on the stored embeddings. ### Generate vector embeddings via PyMilvus and insert them into Zilliz Cloud for semantic search ``` from pymilvus import model, MilvusClient ef = model.dense.VoyageEmbeddingFunction( model_name="voyage-code-2", api_key="your-voyage-api-key", ) # Generate embeddings for documents docs = [ "retriever = KNNRetriever.from_texts(documents, embeddings)", "knn = KNeighborsClassifier(n_neighbors=3)", "sorted_numbers = sorted(numbers)", "def dynamic_programming(): print('yes')", "documents_embds = get_embeddings(documents)", "response = client.embeddings.create(input = documents, model='text-embedding-ada-002')" ] docs_embeddings = ef.encode_documents(docs) # Generate embeddings for queries queries = ["Is the function dynamic_programming() implemented using dynamic programming?"] query_embeddings = ef.encode_queries(queries) # Connect to Zilliz Cloud with Public Endpoint and API Key client = MilvusClient( uri=ZILLIZ_PUBLIC_ENDPOINT, token=ZILLIZ_API_KEY) COLLECTION = "documents" if client.has_collection(collection_name=COLLECTION): client.drop_collection(collection_name=COLLECTION) client.create_collection( collection_name=COLLECTION, dimension=ef.dim, auto_id=True) for doc, embedding in zip(docs, docs_embeddings): client.insert(COLLECTION, {"text": doc, "vector": embedding}) results = client.search( collection_name=COLLECTION, data=query_embeddings, consistency_level="Strong", output_fields=["text"]) ``` For more information, refer to our [PyMilvus Embedding Model documentation](https://milvus.io/docs/embeddings.md).   ### Generate vector embeddings via Voyage AI 's Python SDK and insert them into Zilliz Cloud for semantic search ``` import voyageai from pymilvus import MilvusClient vo = voyageai.Client(api_key="your-voyage-api-key") # Generate embeddings for documents docs = [ "retriever = KNNRetriever.from_texts(documents, embeddings)", "knn = KNeighborsClassifier(n_neighbors=3)", "sorted_numbers = sorted(numbers)", "def dynamic_programming(): print('yes')", "documents_embds = get_embeddings(documents)", "response = client.embeddings.create(input = documents, model='text-embedding-ada-002')" ] docs_embeddings = vo.embed(docs, model="voyage-code-2", input_type="document").embeddings # Generate embeddings for queries queries = ["Is the function dynamic_programming() implemented using dynamic programming?"] query_embeddings = vo.embed(queries, model="voyage-code-2", input_type="query").embeddings # Connect to Zilliz Cloud with Public Endpoint and API Key client = MilvusClient( uri=ZILLIZ_PUBLIC_ENDPOINT, token=ZILLIZ_API_KEY) COLLECTION = "documents" if client.has_collection(collection_name=COLLECTION): client.drop_collection(collection_name=COLLECTION) client.create_collection( collection_name=COLLECTION, dimension=1536, auto_id=True) for doc, embedding in zip(docs, docs_embeddings): client.insert(COLLECTION, {"text": doc, "vector": embedding}) results = client.search( collection_name=COLLECTION, data=query_embeddings, consistency_level="Strong", output_fields=["text"]) ``` For more information, refer to [Voyage AI Embedding Guide](https://docs.voyageai.com/docs/embeddings#python-api).