Posted On: Oct 5, 2023
HAQM Relational Database Service (RDS) for PostgreSQL now supports v0.5.0 of the pgvector extension to store embeddings from machine learning (ML) models in your database and to perform efficient similarity searches. This version of the extension introduces pgvector introduces HNSW indexing support, parallelization of ivfflat index builds, and improves performance of its distance functions.
Embeddings are numerical representations (vectors) created from generative AI that capture the semantic meaning of text input into a large language model (LLM). pgvector can store and search embeddings from HAQM Bedrock, HAQM SageMaker, and more. With pgvector on HAQM RDS, you can simply set up, operate, and scale databases for your GenAI applications. pgvector 0.5.0 adds support for HNSW indexing, which lets you execute similarity searches with low latency and produces highly relevant results. Additionally, HNSW in pgvector supports concurrent inserts, and updating/deleting vectors from the index. You can integrate your GenAI applications with pgvector using open-source frameworks like LangChain, simplifying how you use HAQM RDS for searching over your vector data.
The pgvector extension version 0.5.0 is available on database instances in HAQM RDS running PostgreSQL 15.4-R2 and higher, 14.9-R2 and higher, 13.12-R2 and higher, and 12.16-R2 and higher in all applicable AWS Regions, including the AWS GovCloud (US) Regions.
You can get started by launching a new HAQM RDS DB instance directly from the AWS Console or the AWS CLI. Learn more about pgvector in the AWS Database Blog and the HAQM RDS User Guide.