lightRAG with Google VertexAI embedding models

HKUDS / LightRAG

"LightRAG: Simple and Fast Retrieval-Augmented Generation"

https://arxiv.org/abs/2410.05779

MIT License

9.28k stars 1.14k forks source link

lightRAG with Google VertexAI embedding models #178

Closed anguy044 closed 1 week ago

anguy044 commented 3 weeks ago

I don't believe there's a built in function in lightRAG current for using Google Vertex Embeddings models from looking at the source files (llm.py), correct me if I'm wrong or missed it.

I'm curious if this was something will be available in future updates?

Thank you.

Jaykumaran commented 3 weeks ago

Hello,

Yes current setup doesn't hold support for using Gemini API's. However I think we can run Gemini or Vertex Embeddings via OpenAI Proxy Compatible API. I'm also looking out for this solution. Kindly share if have run the workaround successfully.

https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/call-vertex-using-openai-library

Jaykumaran commented 2 weeks ago

The authors had replied to a similar question in one of previous thread #175 .

You can have a look at this.