Open Source embedding - Nomic

[AltayYuzeir]

Hello. Thank you for open sourcing this awesome tool.
I was watching the video by Matthew Berman and he suggested to improve the open source nature of the project by replacing the OpenAI embedding with the open source Nomic embedding.

[ipty]

you can change the settings in /app/config.tsx
https://github.com/developersdigest/llm-answer-engine/blob/main/app/config.tsx
line 12: embeddingsModel

hope it helps
erik

[developersdigest]

thanks for the rely @ipty , @AltayYuzeir I can confirm that nomic and local embeddings do work. If you are going to use local embedding models while also using a local chat completions, time-to-first-token can increase considerably depending on your machine. I recommend decreasing the values associated with the RAG process at the bottom of the config when using local inference. cheers!