Introducing the Embedding Function: How Milvus 2.6 Streamlines Vectorization and Semantic Search

What this post added

Introduces the Embedding Function (Data-in, Data-out) in Milvus 2.6, a new built-in capability that handles vectorization internally by connecting to external model providers. This feature allows users to insert raw data and query using raw text, with Milvus automatically managing the embedding generation process. It details the 'Function' module framework, the workflow steps, configuration via milvus.yaml, and schema definition including scalar and vector fields, and the embedding function itself. Demonstrates insertion of raw text and performing semantic search using text queries.