GPU is not being used during detection, showing CUDA not available

[spaul13]

I have installed the CUDA 10.1 and corresponding cuDNN 7.5.0 and added to the path variables and everything else according to the documentation

But while running the detection, I check the output of torch.cuda.is_available() which is returning false and also checked the GPU utilization is 0-1% so GPU is not being used.

Can anyone please tell me how to make GPU to be used during detection to make detection faster?