@vicgalle

Hi there,

I'm noticing that the version of the serve.py file on the main branch doesn't seem to actually be creating the zero-shot classification endpoint that should have been added in the v0.3 release. Am I missing something obvious?

I'm super excited to try out that feature!

Thanks,
Daniel

@vicgalle

Hi there,

I'm noticing that the version of the serve.py file on the main branch doesn't seem to actually be creating the zero-shot classification endpoint that should have been added in the v0.3 release. Am I missing something obvious?

I'm super excited to try out that feature!

Thanks,
Daniel

Hi @danielweidman , thank you for your interest in the new version!

I have not made public the changes in the implementation yet since i'm planning on a refactor of the server code (while keeping the public API intact) after I return from vacation.

However, if you are interested in setting up a zero-shot classification model, the one that is currently serving the API is this one: https://huggingface.co/vicgalle/xlm-roberta-large-xnli-anli (there are instructions there).

Let me know if you have further questions!

I see! Thanks very much for the info and that link! I hope your vacation is going well.