A bug when try to modify your code.

Since I'm new to PyTorch, when I add transforms.CenterCrop(size) to get my dataloader, this error occur

  File "~/triplet-network-pytorch/triplet_data_loader.py", line 73, in __getitem__
    img1 = self.transform(img1)
  File "build/bdist.linux-x86_64/egg/torchvision/transforms.py", line 29, in __call__
    img = t(img)
  File "build/bdist.linux-x86_64/egg/torchvision/transforms.py", line 156, in __call__
    w, h = img.size
TypeError: 'builtin_function_or_method' object is not iterable

If I add transforms.CenterCrop(28) to MNIST_t(*),

train_loader = torch.utils.data.DataLoader(
        MNIST_t('./data', train=True, download=True,
                       transform=transforms.Compose([
                           transforms.ToTensor(),
                           transforms.Normalize((0.1307,), (0.3081,)),
                           # transforms.CenterCrop(28)
                       ])),
        batch_size=args.batch_size, shuffle=True, **kwargs)

It also occur such error. Could you please help me solve it?

andreasveit / triplet-network-pytorch

A bug when try to modify your code. #1