Ken Simpson

Just wanted to throw my hat in the ring here. I'm getting the same error when running the image2image workflow from https://comfyanonymous.github.io/ComfyUI_examples/stable_cascade/. Not sure why the encoder would be spitting...

Same error here.

I think the reason Würstchen isn't being taken up by anyone is that the quality is lacking. It's faster to train, certainly, but for all of SD's faults, it has...

I will contribute a PR :)

Some better error checking would be nice throughout the node. I will try to add some more and submit a PR. For instance, if you try to detect something that...

> @ttulttul do you maybe want to join my try to push this repo to dinov2 i would setup a new fork where we can try this. but yesterday i...

Hi Alessandro, I've started following Daniel's fork and I'll see where I can contribute there. Thank you! On Sun, Nov 12, 2023 at 1:18 AM Alessandro Perilli ***@***.***> wrote: >...

Make sure you are using SAMModelLoader (Segment Anything) rather than "SAM Model Loader". Doing so resolved this issue for me. Must be something about how the two model loaders deliver...

Please see PR https://github.com/storyicon/comfyui_segment_anything/pull/16

> The same question, Please You can already do this. The underlying segmentation model allows you to specify separate prompts by putting " . " between prompts. For instance, this...