Possible Causes:
(1) The function name is the same as a name in the system, but this time there is no one named 'parametres'
(2) Import modul or function error
Actual reason:
模型加入了.cuda() is a function, so it cannot be directly optimized
solution:
Remove .cuda and change to(device)
Mirroring problem:
AttributeError: 'Model' object has no attribute 'parameters'
Problem: torch.nn.model is not added to the init in the class
solution:
