Hello everyone, I found a problem when I read the memory address of CNN parameter, I found some problems!
from torchvision import models
resnet = models.resnet50(pretrained=False)
when I try to print the id of the parameters, I found the first print have a lot of memory address, there are dozens of them. But the second print only have one address.
I don’t quite understand this problem, can someone explain it? I mainly used to adjust the learning rate, but I found this strange problem.
I’m not an expert on this. I hope someone can explain.