RuntimeError: cuda runtime error (100) : no CUDA-capable device is detected at /pytorch/aten/src/THC/THCGeneral.cpp:47

nlp
1

hey,
I am getting this error despite trying out all the workarounds available on the forum. The status of my machine and version compatibility seems quite accurate but I still, don’t know why this error was prompted.
Already tried to reinstall PyTorch but nothing works.

Here are the machine validations.

image
image779×446 9.7 KB

@ptrblck could you please assist in this regard to fix this?

2

Your driver seems to be too old, as CUDA 10.2 would need NVIDIA driver >= 440.33.

3

I solved the issue by upgrading the Pytorch from 1.5.0 to 1.6.0 with Cuda 10.1. Also, in the command, I explicitly mentioned the GPU_id=0 as I have just one GPU in my device.