GraphRAG Toolkit

I apologize for the confusion, but there seems to be a misunderstanding regarding the foundation model (FM) name you mentioned. The model "cohere.embed-english-v3" is not listed as an official model name in Amazon Bedrock.

However, Cohere does offer embedding models through Amazon Bedrock. The model you're likely looking for is simply called "Embed" or "Cohere Embed". This model is Cohere's text representation or embeddings model, and it does have an English-only version.

The Cohere Embed model for English has the following characteristics:

• It supports 1024 dimensions
• It is designed for English language use
• It has a maximum token limit of 4K (4096 tokens)
• It is not supported for fine-tuning
• Its primary use cases include semantic search, retrieval augmented generation (RAG), classification, and clustering

To use this model, you would need to refer to it using the appropriate model ID as specified in the Amazon Bedrock documentation. The exact model ID may vary, so it's best to check the current Amazon Bedrock console or documentation for the most up-to-date information on how to access and use the Cohere Embed model.

If you're working with the GraphRAG Toolkit, make sure you have enabled access to the appropriate foundation models in Amazon Bedrock before running your templates or notebooks. You may need to configure the toolkit to use the correct model ID for the Cohere Embed model you want to use.
Sources
Cohere Command & Embed on Amazon Bedrock – AWS
Cohere Command & Embed - Models in Amazon Bedrock - AWS
Introducing the GraphRAG Toolkit | AWS Database Blog
Amazon Bedrock foundation model information - Amazon Bedrock