Solved on its own, please ignore

[xiaoboelse]

Can you help me check this question? Thank you!

[PhilippeWeinzaepfel]

Hi,

I am not familiar at all with cuda/pytorch/compilation on windows.
However, it seems that your installed version of pytorch is a cpu-only build. If you have installed it with conda, do not forget the nvidia channel and the pytorch-cuda, eg conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia.
Then you need to install the cuda toolkit (https://developer.nvidia.com/cuda-downloads) and set CUDA_HOME to the location where it is installed.

Best regards,
Philippe

[xiaoboelse]

ok, Thank you very much for the friendly reminder. It has been installed