Spaces:

---

radames

/

stable-diffusion-depth2img

like 234

Runtime error

hi @aknoerig are you sure? if the seed is set, then the generator = torch.Generator(device=device).manual_seed(seed) is defined with the giving seed right?

hi @radames , yeah, take a look at the params documented here: https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion#diffusers.StableDiffusionDepth2ImgPipeline.__call__
there's no seed param, instead you have to create a generator and provide it via the generator param.
you can actually tell it's currently ignored in your space: even if you set the same seed, it will generate different images.

hehe, just noticed that the build is currently failing due to that exact error. seems it has become less forgiving:
TypeError: __call__() got an unexpected keyword argument 'seed'

oh thanks @aknoerig you're right, diffusers has update their API since I've release the demo. I'll fix it.
Unfortunately there's is a conflict now and I can't merge your commit 🥲

btw @aknoerig thanks for the fix, I've managed the conflict locally and push it here

you're welcome, thank you for making this space available in the first place!