When I train the neural network, I observe that even though I set the device="cuda:0" or "cuda" in the config.py. When I check the process running, the resources it use is still CPU, and there is little to none resource used from GPU. I'm not sure it is normal?
Hi,
When I train the neural network, I observe that even though I set the
device="cuda:0" or "cuda"in theconfig.py. When I check the process running, the resources it use is still CPU, and there is little to none resource used from GPU. I'm not sure it is normal?