Select specific rows of 2D PyTorch tensor

Question

Suppose I have a 2D tensor looking something like this:

[[44, 50, 1, 32],
.
.
.
[7, 13, 90, 83]]

and a list of row indices that I want to select that looks something like this [0, 34, 100, ..., 745]. How can I go through and create a new tensor that contains only the rows whose indices are contained in the array?

Toon Tran · Accepted Answer · 2020-03-22 01:08:01Z

11

You could select like with numpy

import torch
x = torch.Tensor([[1, 2, 3, 4],
                  [5, 6, 7, 8],
                  [9, 8, 7, 6],
                  [5, 4, 2, 1]])

indices = [0, 3]

print(x[indices])
# tensor([[1., 2., 3., 4.],
#         [5., 4., 2., 1.]])

edited Mar 22, 2020 at 1:08

answered Feb 3, 2020 at 2:53

Toon Tran

3785 silver badges8 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Select specific rows of 2D PyTorch tensor

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related