WAS

Comfy can't output anything to the UI except a image, from hardcoded location. Some of it's most crippling issues right now.

> While not core, this extension can help in the meantime : https://github.com/m957ymj75urz/ComfyUI-Custom-Nodes/tree/main/clip-text-encode-split (drop the .py file in your `custom_nodes` folder). This will add 4 nodes : `RawText`, `RawTextCLIPEncode`, `RawTextCombine`...

I'd rather see a fixed portrait size, and just zoom in the landscape scaled images to fit (click to view full size), so the space is completely occupied. Civit AI...

Sounds like it's using Go Big type method, which if so, can't really be overcome without doing stuff like img2img color fix, which also imo, produces bad quality.

That looks spectacular and rather continuous even when raising the levels of the image

Ah drat. Running into so many road blocks.

> I'm working on adding the optional inputs in to the code. > > I actually have the inputs working but I want to make sure that they are visually...

> They should be working now. Awesome! Thanks so much @comfyanonymous

I'm not positive, but I think it's similar to diffusion. Say you got little fish floating around a person with the absence of water, they may come out as birds,...

I have intermittent internet over 4G/5G, and when the internet drops, the WebUI says "Connection lost" error, puts a error over model list (not sure why that would need internet)...