ImportError: cannot import name "CLIPVisionModelWithProjection" from 'transformers'

[doggie015]

Running on windows with the guide at https://www.howtogeek.com/830179/how-to-run-stable-diffusion-on-your-pc-to-generate-ai-images/

Any ideas to fix this error that pops up when I try to run the command "python scripts/txt2img.py --prompt "a close-up portrait of a cat by pablo picasso, vivid, abstract art, colorful, vibrant" --plms --n_iter 5 --n_samples 1"

[qqlu]

I meet a similar problem...

[Steiffj]

I encountered the same problem as well with the latest version of the codebase.

That article was last updated 24 Oct, so I checked out commit 69ae4b3 from 22 Aug and tried again - with this earlier version it worked!

I'm not familiar with the development environment or the code, but it seems like it fails to resolve CLIPVisionModelWithProjection when importing the safety checker. Not sure why though, I didn't see any errors when downloading the environment packages.

[doggie015]

I encountered the same problem as well with the latest version of the codebase.

That article was last updated 24 Oct, so I checked out commit 69ae4b3 from 22 Aug and tried again - with this earlier version it worked!

I'm not familiar with the development environment or the code, but it seems like it fails to resolve CLIPVisionModelWithProjection when importing the safety checker. Not sure why though, I didn't see any errors when downloading the environment packages.

That worked a treat. Now the error is CUDA out of VRAM but that tracks, I have a 1650 4GB graphics card so it's below the requirements. I plan to upgrade in the future.