R
RunPod6mo ago
Kiwino

NansException: A tensor with all NaNs was produced in Unet.

hi,after i upload a model from huggingface ,my sd cant generate anymore ,it says "NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check."what should i do, i try to add "--no-half"after the "set COMMANDLINE_ARGS=" in webui-user.bat .And i also cross theUpcast cross attention layer to float32,but it doesnt work still.
2 Replies
nerdylive
nerdylive6mo ago
This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check Maybe its because ur model
digigoblin
digigoblin6mo ago
You added it in the wrong place .bat files are for Windows not Linux. Add to webui-user.sh not .bat
Want results from more Discord servers?
Add your server