ride5k
ride5k
> In the following comparisons, "DPM++ 2M Test" is the modified function posted earlier in the thread, the loss of detail is extremely obvious. "DPM++ 2M TestX" is an altered...
came here to suggest the same as well
> line 124 in txt2mask.py: > > `model.load_state_dict(torch.load(d64_file, map_location=torch.device('cuda')), strict=False);` > > change `cuda` to `cpu` this worked for me too... kudos!
can confirm same issue on rx6700xt using zluda EDIT: carefully re-reading, i am NOT experiencing the issue when using inpaint or inpaint global harmony, however issue does appear when using...
> What type of hardware are you using? Is it an intel or AMD graphics card? @huchenlei i am running the directml fork on AMD.
fyi I am able to use with GPU processing on amd 6700xt via zluda on a1111.
having same issue of persistent text block. extension list cut and pasted from box: ,txt,img tabs,left,right default,left,right GFPGAN,left,right CodeFormer,left,right ADetailer,left,right Tiled Diffusion,left,right DemoFusion,left,right Tiled VAE,left,right Face Editor,left,right Aspect Ratio Helper,left,right...
> > "--disable-features=CalculateNativeWinOcclusion" works . (Chrome 134) > > what does that mean? https://www.reddit.com/r/chrome/comments/rxoduk/comment/hwk00hw/
> I found the bug! fantastic work. i wiped my venv today and upon rebuild got the same error messages. substituting "pretrained_model_name_or_path" in place of None worked perfectly!
would love a fix for this as well!