# How to shuffle columns or rows of matrix in PyTorch?

PyTorchServer Side ProgrammingProgramming

#### Complete Python Prime Pack

9 Courses     2 eBooks

#### Artificial Intelligence & Machine Learning Prime Pack

6 Courses     1 eBooks

#### Java Prime Pack

9 Courses     2 eBooks

A matrix in PyTorch is a 2-dimension tensor having elements of the same dtype. We can shuffle a row by another row and a column by another column. To shuffle rows or columns, we can use simple slicing and indexing as we do in Numpy.

• If we want to shuffle rows, then we do slicing in the row indices.

• To shuffle columns, we do slicing in the column indices.

• For example, if we want to shuffle the 1st and 2nd rows of a 3☓3 matrix, then we just shuffle the index of these rows and make a slicing to find the shuffled matrix.

Let's take a couple of examples to have a better understanding of how it works.

## Steps

You could use the following steps to shuffle the rows or columns of a matrix.

• Import the required library. In all the following examples, the required Python library is torch. Make sure you have already installed it.

import torch
• Define a matrix and print it.

matrix = torch.tensor([[1., 2., 3.],[4., 5., 6.],[7, 8, 9]])
print("Original Matrix:", matrix)
• Specify the row and column indices with shuffled indices. In the following example we shuffle 1st and 2nd row. So, we interchanged the indices of these rows.

# shuffle 1st and second row
r = torch.tensor([1, 0, 2])
c = torch.tensor([0, 1, 2])
• Shuffle the rows or columns of the matrix.

matrix=matrix[r[:, None], c] # shuffles rows
matrix = matrix[r][:,c] # shuffles columns
• Print the shuffled matrix.

print("Shuffled Matrix:", matrix)

## Example 1

In the following example, we shuffle the 1st and 2nd rows.

# Import the required library
import torch

# create a matrix
matrix = torch.tensor([[1., 2., 3.],[4., 5., 6.],[7, 8, 9]])

# print matrix
print("Original Matrix:", matrix)

# shuffle 1st and second row
r = torch.tensor([1, 0, 2])
c = torch.tensor([0, 1, 2])
matrix=matrix[r[:, None], c]
print("Shuffled Matrix:", matrix)

## Output

Original Matrix:
tensor([[1., 2., 3.],
[4., 5., 6.],
[7., 8., 9.]])
Shuffled Matrix:
tensor([[4., 5., 6.],
[1., 2., 3.],
[7., 8., 9.]])

Notice that the 1st and 2nd rows are shuffled.

## Example 2

In the following example, we shuffle 2nd and 3rd columns.

# Import the required library
import torch

# create a matrix
matrix = torch.tensor([[1., 2., 3.],[4., 5., 6.],[7, 8, 9]])

# print matrix
print("Original Matrix:", matrix)

# shuffle 2nd and 3rd columns
r = torch.tensor([0, 1, 2])
c = torch.tensor([0, 2, 1])
matrix_col_Shuffled = matrix[r][:,c]
print("Shuffled Matrix:", matrix_col_Shuffled)

## Output

Original Matrix:
tensor([[1., 2., 3.],
[4., 5., 6.],
[7., 8., 9.]])
Shuffled Matrix:
tensor([[1., 3., 2.],
[4., 6., 5.],
[7., 9., 8.]])
Updated on 20-Jan-2022 07:06:23

Advertisements