-
Notifications
You must be signed in to change notification settings - Fork 12
RuntimeError: mat1 and mat2 must have the same dtype #11
Comments
Try removing --no-half or --no-half-vae from cmdline args |
yep it works but I get black images ^^' |
You're on AMD hardware? |
Remove --disable-nan-check,and try again |
No , Nvidia GPU and intel core |
didn't work sadly ... |
Thats odd you shouldn't get NaNs with Nvidia hardware, which GPU? |
I have a RTX 3060 , and no I didn't. I made a fresh install but now I get the error OutOfMemoryError: CUDA out of memory. Tried to allocate 26.00 MiB (GPU 0; 12.00 GiB total capacity; 11.06 GiB already allocated; 0 bytes free; 11.30 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF I didn't even get it before ... |
It does work now, but only at 1024X1024, not more , but that's regardless of the use of the extension ... |
Try --medvram |
Hello,
I'm trying to use the extension but I get this error when the refiner kicks in. Any idea ?
RuntimeError: mat1 and mat2 must have the same dtype
Best regards !
The text was updated successfully, but these errors were encountered: