Spaces:

Hansimov

/

hf-llm-api

like 13

Running

Currently this project would only support the models available in https://huggingface.co/chat.
No plan for any model that has no inference endpoint.

BTW, I have dived into your forked spaces and noticed some trial and errors.
Remember to upgrade to the latest version of this project.

Ok got it. Can you add these models NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO
codellama/CodeLlama-70b-Instruct-hf.

Hi @MAsad789565 ,
NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO is already supported, you can pass model="nous-mixtral-8x7b" to use it.

I have previously supported codellama/CodeLlama-34b-Instruct-hf, but it behaves far below expected, so I disabled it.
I have not yet tested the new 70b version. So if you think it is better enough and provide some cases, I might consider to enable it again.

You can also go to this GitHub issue, I have listed all the supported models and why some models are not supported there:

• [Feature Request] Support all models provided by HuggingChat
  • https://github.com/Hansimov/hf-llm-api/issues/5