PyTorch .hsplit()

In PyTorch, the .hsplit() method splits a tensor into multiple sub-tensors horizontally (column-wise) along the specified dimension (axis).

Syntax

torch.hsplit(tensor, indices_or_sections)

• tensor: The tensor to be split.
• indices_or_sections: This can be an integer or a list of integers.
  • If it’s an integer, it specifies the number of equal-sized sub-tensors to split the tensor into.
  • If it’s a list of integers, it specifies the sizes of each sub-tensor along the specified dimension.

Example

The following example demonstrates the usage of the .hsplit() method:

import torch

# Define a tensor
tensor = torch.tensor([[1, 2, 3, 4],
                       [5, 6, 7, 8],
                       [9, 10, 11, 12]])

# Split the tensor into 2 equal parts
sub_tensors = torch.hsplit(tensor, 2)

print(sub_tensors)

Copy to clipboard
Copy to clipboard

The above code produces the following output:

(tensor([[ 1,  2],
        [ 5,  6],
        [ 9, 10]]), tensor([[ 3,  4],
        [ 7,  8],
        [11, 12]]))

Copy to clipboard
Copy to clipboard

The tensor is split into two sub-tensors along the columns. The first sub-tensor contains the first two columns of the original tensor, while the second sub-tensor contains the last two columns.