Skip to content

PyTorch error in lab three #18

Description

@cxzhou35

When I am doing lab three, I met the error: module 'torch.nn' has no attribute 'Buffer' in the class below:

class MNISTSampler(nn.Module, Sampleable):
    """
    Sampleable wrapper for the MNIST dataset
    """
    def __init__(self):
        super().__init__()
        self.dataset = datasets.MNIST(
            root='./data',
            train=True,
            download=True,
            transform=transforms.Compose([
                transforms.Resize((32, 32)),
                transforms.ToTensor(),
                transforms.Normalize((0.5,), (0.5,)),
            ])
        )
        self.dummy = nn.Buffer(
            torch.zeros(1)
        )  # Will automatically be moved when self.to(...) is called...

My PyTorch version is 2.4.1+cu121.
I try to instead with the self.register_buffer('dummy', torch.zeros(1)), and it works.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions