✨ Add inferenceProvider filter when listing models

[frascuchon]

This PR adds the missing query param inference_provider when listing models (listModels).

[julien-c]

let's wait for a @SBrandeis review to make sure the API is final

[Wauplin]

(following this PR to implement the same in huggingface_hub once settled)

[SBrandeis]

You can search for several providers at once

Suggested change

	inferenceProvider?: string;
	inferenceProvider?: string[];

[frascuchon]

Passing more than one provider raises an error:

Error: Internal Error - We're working hard to fix this as soon as possible!. URL: https://huggingface.co/api/models?limit=10&expand=pipeline_tag&expand=private&expand=gated&expand=downloads&expand=likes&expand=lastModified&expand=inferenceProviderMapping&inference_provider=together&inference_provider=replicate. Request ID: Root=1-67ae03c6-377f36e3001096932c7699a4

[SBrandeis]

With this URL it works:
https://huggingface.co/api/models?limit=10&expand=pipeline_tag&expand=private&expand=gated&expand=downloads&expand=likes&expand=lastModified&expand=inferenceProviderMapping&inference_provider=together,replicate

inference_provider=together,replicate instead of inference_provider=together&inference_provider=replicate

That being said - this should not end up in a Server Error 😵 💫

[frascuchon]

Oh cool. Thanks!. I thought the query param had the same syntax as the expand one

[frascuchon]

And the filter terms are combined with an OR clause, right? I mean inference_provider=together,replicate returns models with one of those providers.

[julien-c]

yes it's an OR

[frascuchon]

Okay. I think it's ready.

[coyotte508]

Thank you @frascuchon !