How do I pass a list of tensors to my network?

vision
1

I am training a Variational Autoencoder. I have a dataset of RGB images with resolution 256*256

images = torchvision.datasets.ImageFolder(root = 'data/', transform=transforms.Compose([transforms.ToTensor()]))
data_loader = torch.utils.data.DataLoader(images, batch_size=32, shuffle=True)

However, my batches are turned into a list of tensors. As a result I cannot pass it to my model for training.

What kind of adjustment should I do in order for my model to be able to train? Thanks in advance

2

The default collate function used in the DataLoader should create tensors from all passed samples, so I’m unsure why a list is returned unless you are using a custom collate_fn.
Which PyTorch version are you using and could you post the folder structure, please?