Use sparse.FloatTensor as DataLoader input

autograd
1

Hi

I would like to use sparse.FloatTensor as DataLoader input.

import torch
from torch.utils.data import TensorDataset, DataLoader

X = torch.sparse.FloatTensor(torch.LongTensor([[0, 1, 1], [2, 0, 2]]), torch.FloatTensor([3, 4, 5]),torch.Size([2,3]))
y = torch.Tensor([0,1])
dataSet = TensorDataset(X, y)
loader = DataLoader(dataSet)
for d, t in loader:
    print(d)

But this code outputs RuntimeError “sparse tensors do not have strides”.
How to fix it?

Thanks.