search

ahrm / UnstableFusion

A Stable Diffusion desktop frontend with inpainting, img2img and more!
GNU General Public License v3.0
1.25k stars 86 forks source link

AssertionError: Torch not compiled with CUDA enabled #15

Closed SnekCode closed 1 year ago

SnekCode commented 1 year ago

Hello,

Installed all dependencies using pip install -r requirements.txt but receive this error: AssertionError: Torch not compiled with CUDA enabled

I assume line 8 torch>=1.12.1 in the requirement.txt is incorrect and should be something like touch>=1.12.1=py38_cu113. Haven't tested this yet nor I'm I confident those are the correct versions for a cuda compiled version of torch. Just posting incase you have a quicker answer!

ahrm commented 1 year ago

Yes, however, if you don't want to run the server locally, you don't need the CUDA dependencies, that's why I didn't include it in the requirements.

SnekCode commented 1 year ago

Ahh ok that makes sense. For the complete SD noobs like myself would it help to add that distinction to the readme? And for documentation purposes for folks that want to run locally they should run pip install torch==1.12.1+cu113 -f https://download.pytorch.org/whl/torch_stable.html?

ahrm commented 1 year ago

Added some documentation in the README.

SnekCode commented 1 year ago

Added some documentation in the README.

Ahh you beat me to it! Thanks!