AssertionError: CUDA unavailable, invalid device 0 requested

In the use of yolov5 During training , Prompt this error , see torch.cuda.is_available() The return value is false

see nvidia-smi Information can be seen

Due to previous use conda Installed torch1.9.0

Use here pip reinstall ：

pip3 install torch == 1.9.0+cu111 torchvision == 0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html

After installation , Re execution torch.cuda.is_available() when

You can see the correct result

Reference resources ：AssertionError: CUDA unavailable, invalid device 0 requested_ eliminate BUG Duck blog -CSDN Blog _cuda unavailable