comfyanonymous

https://github.com/comfyanonymous/ComfyUI/commit/58812ab8ca601cc2dd9dbe64c1f3ffd4929fd0ca That new model should work now. Just a note that this one needs clip_l instead of clip_h, you can download clip_l from: https://huggingface.co/IDKiro/sdxs-512-dreamshaper/blob/main/text_encoder/model.safetensors but other than that the above...

 I'm getting some success sampling this with euler ancestral while adjusting the eta of the sampler. The workflow should be attached to the image.

Assuming I don't find any issues I'll merge this along with #2146 in a few days after I get back home.

That's normal increasing batch size can change a tiny bit the results due to how some algorithms are implemented. You can try running comfyui with --deterministic which might solve your...

Try downloading a fresh standalone package and try it with that.

Does anyone else have this issue? If not it might be a hardware/driver issue.

Can you implement the percentage with the percent to sigma functions like this? https://github.com/comfyanonymous/ComfyUI/blob/master/comfy_extras/nodes_model_downscale.py#L23

It's implemented in the main repo now, you can use this workflow until I write an examples page for it. https://gist.github.com/comfyanonymous/0f09119a342d0dd825bb2d99d19b781c

https://github.com/comfyanonymous/ComfyUI/issues/2785#issuecomment-1949967415 This should be fixed now.

Sorry about that. I was planning on adding some more image processing nodes and properly importing either opencv or kornia and getting rid of that copy paste but never got...