Training "never finishes" or system crashes using PyTorch - GPU has memory allocated but always has 0% utilization using DataLoader

Yes, it is natural that you cannot reproduce the problem because it is already clear that the problem is in my environment, this has already been observed. Well, it’s too bad, so does that mean you won’t be able to try to help me anymore? Remember that this behavior only occurs if the code uses the GPU.

Related (post in this forum): Cpu cores not working.