NVIDIA’s website says cuda is enabled on M620 GPU, but i have not been able to get torch.cuda.is_available() to return True. I notice that when I try download pytorch using: ‘conda install pytorch torchvision cudatoolkit=10.1 -c pytorch’ the build of pytorch that is downloaded is py3.7_cpu_0 which I think may be the problem?
I am using windows 10 and python 3.7
Maybe your NVIDIA driver needs to be updated.
I updated them according to this page (https://www.nvidia.com/content/DriverDownload-March2009/confirmation.php?url=%2FWindows%2FQuadro_Certified%2F440.97%2F440.97-quadro-desktop-notebook-win10-64bit-international-whql.exe&lang=us&type=Quadro&aliId=eyJpIjoiVzVPKzQ1VkJ6VGpHM3pxVSIsInQiOiJWdUpuUFZcL2Q2SkZUZ21FbFNTM0l6dz09In0%3D)
but still no luck getting cuda enabled unfortunately. Wondering why my install of pytorch defaults to a cpu version?
Is nvidia-smi detecting the card correctly and shows the right driver version?
I think it may be seeing my first GPU not my second one?
I updated the drivers recently