Currently, only the OpenAI API seems to be readily supported, and integrating LiteLLM might prove to be more useful with regards to using Huggingface models (such as https://huggingface.co/tiiuae/falcon-40b-instruct, which is released under the Apache 2.0 license, and for which the community provides guidance for finetuning).
Currently, only the OpenAI API seems to be readily supported, and integrating LiteLLM might prove to be more useful with regards to using Huggingface models (such as https://huggingface.co/tiiuae/falcon-40b-instruct, which is released under the Apache 2.0 license, and for which the community provides guidance for finetuning).