ClashSAN

@saltyollpheist no, this user is using nvidia, see "CUDA" reports should be separated to the direct-ml fork as memory doesn't decrease after use for many ml tasks but this isn't...

Hi @sayakpaul, @patrickvonplaten, appreciate and thank you for your input! There are some third party extensions for hf downloads: (not caching-related) https://github.com/camenduru/stable-diffusion-webui-huggingface https://github.com/etherealxx/batchlinks-webui Q: Do other webuis cache the stable-diffusion...

great, thanks! the caching could be useful, if users are switching models mid-sampling as experiments, with deforum and other extensions too. fortunately or unfortunately, AUTOMATIC1111 doesn't always read from the...

you need 16gb of ram to run, or if you have 8gb, make a pagefile with at least 8gb your log is showing MemoryError

reopen if needed

@Centurion-Rome do you have some pictures comparison?

ok this i guess is comparison: 

@catboxanon oh, very cool, is it this? https://huggingface.co/Xynon/noise-offset/tree/main I tried with the model above, the effect would produce darker atmospheres and lighting ~~particularly very prevalent at step 8-10, DPM++ 2M...

you can play with merging if you like https://huggingface.co/ClashSAN/pyramid_noise_test_600steps_08discount/tree/main is this issue resolved?

see if you can manually install C43H66N12O12S2's build of xformers, the recent xformers version breaks training. or just remove xformers from args entirely, see if that works.