python-pytorchHow can I use the Softmax function in Python with PyTorch?

The Softmax function is a popular activation function used in many machine learning applications. It can be used in Python with PyTorch to calculate the probabilities for each class in a multi-class classification problem.

To use the Softmax function in Python with PyTorch, we can use the torch.nn.functional.softmax function. This function takes an input tensor of shape (N, C) and returns a tensor of shape (N, C) containing the softmax values of the input tensor.

Here is an example of how to use the torch.nn.functional.softmax function:

import torch

# Create a tensor of shape (2, 3)
input_tensor = torch.randn(2, 3)

# Calculate the softmax values of the input tensor
softmax_tensor = torch.nn.functional.softmax(input_tensor, dim=1)

print(softmax_tensor)

Output example

tensor([[0.3096, 0.3790, 0.3114],
        [0.4536, 0.2107, 0.3357]])

The code above performs the following steps:

1. Import the torch module.
2. Create an input tensor of shape (2, 3).
3. Calculate the softmax values of the input tensor using the torch.nn.functional.softmax function.
4. Print the resulting softmax tensor.

For more information about the Softmax function in PyTorch, please refer to the PyTorch documentation.

Edit this code on GitHub