.tile()

In PyTorch, the torch.tile() function generates a new tensor by repeating the elements of an existing tensor along specified dimensions. This function is available in the torch module and allows element-wise repetition along multiple dimensions.

Syntax

torch.tile(input, dims)

• input: The input tensor to be repeated.
• dims: A tuple specifying the number of times the tensor is to be repeated along each dimension. The length of dims must match the number of dimensions in input.

Example

The following example demonstrates the usage of the .tile() function:

import torch

# Define a tensor
ten = torch.tensor([13, 24, 35, 46])

# Repeat the tensor 2 times along the first dimension and 3 times along the second dimension
res = torch.tile(ten, (2, 3))

# Print the result
print(res)

The above code produces the following output:

tensor([[13, 24, 35, 46, 13, 24, 35, 46, 13, 24, 35, 46],
        [13, 24, 35, 46, 13, 24, 35, 46, 13, 24, 35, 46]])