Can we use the matrix factorization model locally by downloading it to our local?

[nileshm11]

I am using a matrix factorization (mf) router in a RAG application and want to download the MF model to my local system. Is this possible? Could you also explain how it works internally? Additionally, can we change the embedding model used by the MF router?

[iojw]

Yes, it is possible to download this from HuggingFace! https://huggingface.co/routellm/mf_gpt4_augmented. However, note that you will still have to generate embeddings for OpenAI as input for the time being - @thwu1 is working on adding support for other embedding models in #17!

Please refer to the paper for specific details for how it works: https://arxiv.org/abs/2406.18665

[nileshm11]

Could you please provide more details or share the implementation code for integrating this into a local system? Specifically, I'm interested in how to set up and use the model, including any necessary configurations and dependencies.