I have followed the installation instructions from the website: https://pytorch.org/get-started/previous-versions/ to install PyTorch v1.9.0 with CUDAtoolkit of v11.3.0 on Windows11. However, when running the code, it always shown that ‘Torch is not compatible with CUDA enabled’. I have already installed CUDA v11.3 on Windows11 as well. My device is: Nvidia RTX 3090Ti, Windows11 OS.
Can I know if PyTorch v1.9.0 is compatible with CUDA 11.3? How can I solve this issue?
I don’t remember with CUDA version was used for the
1.9.0 release, but what did the install log show?
Did it show the proper wheel name with the
cu113 tag in it or, in case you are using conda, did it install
Hi @ptrblck ,
Thanks for your reply. According to the official installation instruction (https://pytorch.org/get-started/previous-versions/),
1.9.0 should be compatible with CUDA 10.2 and 11.3. I used
conda to install it using the official installation command:
conda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cudatoolkit=11.3 -c pytorch -c conda-forge.
Do you think if I need to use
pip to install it instead?