Amazon Titan Text Embeddings V2 now available in Amazon Bedrock, optimized for improving RAG

This article announces the availability of Amazon Titan Text Embeddings V2, a new and improved text embeddings model optimized for Retrieval-Augmented Generation (RAG) in Amazon Bedrock. It provides an overview of the new features and capabilities of this model.

Specifically, the article covers:

• Key features of Amazon Titan Text Embeddings V2, including support for smaller vector sizes (256, 512, or 1024 dimensions) and improved vector normalization for better accuracy
• How embeddings help improve the accuracy of RAG by providing condensed summaries of text to retrieve the most relevant information for language models
• Overview of the model's specifications, such as maximum tokens, supported languages, and vector size options
• An example code snippet in Swift for invoking the model using the Amazon Bedrock Runtime API
• Information on availability, pricing, and additional resources for learning more about Amazon Titan models in Amazon Bedrock