I want to calculate the differences between adjacent elements of a tensor X
in 1 D.
e.g if my tensor has a dimension of BxC
where B
is batch size and C
is a channel.
function diff(X)
should outputt [X(2)-X(1) X(3)-X(2) ... X(c)-X(c-1)]
.
How can I do it in pytorch?
This code should work to compute the difference in the C
dimension:
B, C = 10, 5
x = torch.arange(B * C).view(B, C)
result = x[:, 1:] - x[:, :-1]
2 Likes