Weaviate 1.24 Release: Named Vectors and BQ Improvements
AI Impact Summary
Weaviate 1.24 introduces a significant enhancement: named vectors, allowing objects within collections to have multiple, independent vector embeddings. This unlocks richer querying capabilities, enabling users to vectorize metadata like titles and descriptions and group related objects based on these vectors. The improved binary quantization (BQ) further accelerates vector search by dramatically reducing storage requirements and improving query speed, particularly when combined with the HNSW index.
Affected Systems
Business Impact
Teams can now leverage more granular vector search, improving data retrieval accuracy and query performance, particularly for complex datasets with diverse metadata.
- Date
- Date not specified
- Change type
- capability
- Severity
- info