Batch-wise reordering of a tensor to image

2

Hi,

I think this link is what you want.
In this case, I try the following code:

input = torch.tensor([[[1,2],
                       [3,4]],
                      [[5,6],
                       [7,8]],
                      [[9,10],
                       [11,12]],
                      [[13,14],
                       [15,16]]])
input = input.view(1,4,4) # 1 is used to fix the dim needed
output = input.unfold(2,2,2).unfold(3,2,2)
output = output.contiguous().view(4,4)