python-pytorchHow do I flatten a tensor in Python using PyTorch?

To flatten a tensor in Python using PyTorch, you can use the .view() method. This method reshapes a tensor into a different size without changing its data.

For example:

import torch

x = torch.randn(2,3,4)

y = x.view(-1)

print(y)

Output example

tensor([ 0.1772, -0.0717,  0.0362, -1.0058, -0.3590, -0.3914,  0.7644,  0.9333,
        -0.3182,  1.0144,  1.9072,  0.5056])

The code above creates a tensor with shape (2,3,4) and then reshapes it into a tensor with shape (12,) using the .view() method.

The .view() method takes in two parameters:

• -1 which means that the size of the dimension should be inferred from the size of the tensor.
• size which is a tuple that specifies the desired shape of the tensor.

For more information on the .view() method, please refer to the PyTorch Documentation.

Edit this code on GitHub