private-gpt icon indicating copy to clipboard operation
private-gpt copied to clipboard

Add customizable embedding model support for OpenAI API

Open rajandangi opened this issue 1 year ago • 1 comments

This pull request enables the specification of embedding models in the OpenAI settings. Changes include updates to configuration files and the embedding component.

Changes Made

  • Modified settings-openai.yaml to accept embedding_model under the llm section.
  • Updated settings.yaml to define embedding_model for the OpenAI service.
  • Altered embedding_component.py in the private_gpt to initialize OpenAIEmbedding with a user-defined model.
  • Added embedding_model field in private_gpt/settings/settings.py with a default value and description.

This enhancement allows for flexible selection of OpenAI embedding models, enhancing adaptability to various use cases.

rajandangi avatar May 02 '24 13:05 rajandangi

Covers the same functionality as https://github.com/zylon-ai/private-gpt/pull/1920.

imartinez avatar May 10 '24 14:05 imartinez