I have the following code:
x = (torch.arange(16)+1).reshape(1,1,4,4).float()
im2col = F.unfold(x, (2,2))
with the x
being:
1 2 3 4
5 6 7 8
9 10 11 12
13 14 15 16
I expect the this to have an output of
[[1,2,5,6],
[2,3,6,7],
[3,4,7,8],
.....]
but it is giving me the output of kernel size 3:
[[1,2,3,5,6,7,9,10,11],
[2,3,4,6,7,8,10,11,12],
.....]
Am I thinking wrong about the unfold operation? and is this the expected behavior? When so, can you elaborate?