Question

#16
by deleted - opened
deleted

I am curious of how hard is to redo all of your models to make them work with updated API Inference for your spaces. Sorry if this is a stupid question. I miss your awesome models.

Owner

Just show me a space that does it and I can copy the code from them! :D

deleted

I meant how hard is it to redo all 505 models you created to make them work on your spaces for the updated API inference. Sorry for the confusion.

Owner

If I knew how to do that, it'd be a start, I meant that if you've seen someone redo a single model so it works for the update API inference I could check what they did and copy them, and then I'd know how hard is it. I haven't seen anybody making any model work yet.

I saw someone put a request in to get the inference for john6666, yntec and digiplay and no one even replied.

https://huggingface.co/spaces/huggingface/InferenceSupport/discussions/3977

Owner

We only need a single model being supported by the Inference API, say, this one: https://huggingface.co/Yntec/stable-diffusion-v1-5 The rest can be turned into Loras and used that way as adapters, the way Flux is doing it. Though, I'm not even sure that would work. If nothing ever comes back, we should feel lucky that we were here to live the dream while it lasted.

Sign up or log in to comment