I'm giving comfy UI a shot and it seems ridiculously faster than automatic at loading and generating
Stable Diffusion
Discuss matters related to our favourite AI Art generation technology
Also see
Other communities
Just tried it for the first time, crashed my PC. 10/10 would recommend.
I've got it running on a RTX2060. It loads up fairly fast but after a switching models a couple of times I sometimes get out of memory errors, regardless of the image size. The only solution is to kill the process and restart it. There's probably a memory leak somewhere. Generations at 1024x1024 and with 20 steps take around 10 seconds on my machine.
It's so cute!
Same here, it uses way too much RAM. Seems like there's a memory leak.
Generated a lot of photos today, I'm on 32gb ram, 8gb vram 1080. Takes awhile to load when starting auto1111, stalls in the first image gen, then does ok. I notice that at 95% it stalls again, but then spits out the images. It's slow but the quality is great.
Sadly, wait for a1111 updates or use comfyui
Yeah, just wasn't sure if it was my system, my setup, or sdxl. Since automatic is so finicky.