Hi, I have a question about this code?

[BCWang93]

When I run this code FCN32s to training voc2012 dataset,I get some error like this: the origin image size is like this: torch.Size([1, 3, 333, 500])#image torch.Size([1, 333, 500])#label but the out size of the net is this: torch.Size([1, 21, 141, 301])#output of net So,I think this have some problem,the size of (h,w) is not same as the origin!Can you help me to solve this?Thanks!