Amazon Bedrock now have support for Amazon Aurora PostgreSQL and Cohere embedding models
Table of contents
No headings in the article.
AWS has introduced additional capabilities and upgrades to Knowledge Bases for Amazon Bedrock, with the goal of giving users greater freedom and customization choices. One major innovation is the inclusion of Amazon Aurora PostgreSQL as a custom vector store option, which allows customers to store, index, and search for vector embeddings without updating current apps or tools. Additionally, two new embedding models, Cohere Embed English and Cohere Embed Multilingual, have been added to the current Amazon Titan Text Embeddings, which convert data into vector embeddings.
The Amazon OpenSearch Serverless interface has been enhanced to include Pinecone serverless as a custom vector store option, which reduces development and testing expenses. Users may now select either Pinecone or Pinecone serverless setups and disable extra replicas for development and testing. Another new feature for foundation models (FMs) is Anthropic Claude 2.1, which allows for a 200 K token context window. These upgrades are now available in the US East (North Virginia) and West (Oregon) areas.
In summary, these enhancements are intended to give users additional flexibility, customization choices, and cost savings when using Knowledge Bases for Amazon Bedrock.
To read on this topic with more details & in-depth visit: https://www.linkedin.com/posts/imsampro_aws-amazonbedrock-knowledgebases-activity-7163479004379901952-rlez/
Thanks for reading!