Output and Broadcast shape mismatch in MNIST, torchvision

问题

I am getting following error when using MNIST dataset in Torchvision

RuntimeError: output with shape [1, 28, 28] doesn't match the broadcast shape [3, 28, 28]

Here is my code:

import torch
from torchvision import datasets, transforms

transform = transforms.Compose([transforms.ToTensor(),
                            transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
                          ])
trainset = datasets.MNIST('~/.pytorch/MNIST_data/', download=True, train=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)
images, labels = next(iter(trainloader))

回答1:

The error is due to color vs grayscale on the dataset, the dataset is grayscale.

I fixed it by changing transform to

transform = transforms.Compose([transforms.ToTensor(),
  transforms.Normalize((0.5,), (0.5,))
])

来源：https://stackoverflow.com/questions/55124407/output-and-broadcast-shape-mismatch-in-mnist-torchvision