JamAIBase icon indicating copy to clipboard operation
JamAIBase copied to clipboard

Infinity: large update, bump to v2, new models

Open michaelfeil opened this issue 4 months ago • 0 comments

This pull request includes updates to several configuration files to change the models used for embedding and re-ranking, as well as an update to the Docker image version. The most important changes include updates to the .env file, the CI workflow, the Docker Compose configuration, and the model configurations.

Model Updates:

  • .env: Updated EMBEDDING_MODEL to BAAI/bge-small-en-v1.5 and RERANKER_MODEL to mixedbread-ai/mxbai-rerank-xsmall-v1.
  • .github/workflows/ci.yml: Updated EMBEDDING_MODEL and RERANKER_MODEL to match the new values in the .env file.
  • services/api/src/owl/configs/models.json: Updated model IDs and context lengths for embedding and re-ranking models. [1] [2]

Docker Image Update:

  • docker/compose.cpu.yml: Updated the Docker image for the infinity service to version 0.0.63.

michaelfeil avatar Oct 11 '24 06:10 michaelfeil