hello im using cuda 12.5 with cudann v8.9.7 but my device is unable to detect my 1650 gpu i have set the path and everything how to fix it?
which version of pytorch did you install? Is it 2.4.1? I tried everything you did including restart, it did not work still…
- I have the same issue. I used your approach to finally see
torch.cuda.is_available()
returnTrue
for CUDA 12.1 - My GPU is RTX 4070, CUDA 12.6