RuntimeError: Attempting to deserialize object on CUDA device 1 but torch. cuda. device_ count() is 1.

Successfully solved
RuntimeError: Attempting to deserialize object on CUDA device 1 but torch.cuda.device_count() is 1.

Wrong content

The program reports an error at this step
checkpoint = torch.load(‘model5_4.pt’)

The above problem description means that we have not obtained the information about cuda, Because my model runs on the server , The environment is different after downloading to the local .

resolvent

If you are currently only CPU Running in the environment , Need to add map_location=torch.device(‘cpu’).
If you currently have CUDA Running in the environment , Need to add map_location=torch.device(‘cuda’).
checkpoint = torch.load(‘model5_4.pt’)
Replace with ：

checkpoint = torch.load('model5_4.pt',map_location='cuda')

The successful running ！ It's wrong not to report ！