Disable the exllama on all non-cuda devices.

[yuanwu2017]

What does this PR do?

Fixes # (issue)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

Who can review?

[yao-matrix]

@IlyasMoutawwakil , could you help review? This PR is to disable exllama on non-cuda devices, since other devices don't support it yet, this will enable devices like intel habana can work w/ autogptq. The one open is: do we need add rocm into SUPPORT_EXLLAMA_DEVICES, since the exllama says "it's theoretically supported but not validated", could you share your insights?

[IlyasMoutawwakil]

@yao-matrix rocm is registered as cuda so that shouldn't be an issue, I think there's no need for a list of supported devices as exllama is written in cuda and will probably on work on cuda/hip/rocm machines, that should simplify the PR.

[SunMarc]

Thanks for the PR @yuanwu2017 ! LGTM !

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[IlyasMoutawwakil]

@yao-matrix can you run make style please

[yuanwu2017]

@IlyasMoutawwakil The error is not related with the patch. Please help to merge it.