PyTorch .cos()

The .cos() function computes the cosine of each element in the input tensor, applied element-wise, and returns a tensor of the same shape. It’s part of PyTorch’s math operations used in scientific computing and deep learning.

Syntax

torch.cos(input, *, out=None) → Tensor

Copy to clipboard
Copy to clipboard

Parameters:

• input (Tensor): Input tensor with elements in radians.
• out (Tensor, optional): Optional tensor to store the output.

Return value:

A tensor with the cosine of each element in the input, having the same shape.

Example 1: Using .cos() with a 1D tensor

In this example, .cos() computes the cosine of a 1D tensor containing angles in radians:

import torch

# Create a tensor with values in radians
x = torch.tensor([0, torch.pi / 2, torch.pi])

# Compute the cosine
y = torch.cos(x)

print(y)

Copy to clipboard
Copy to clipboard

The output of this code is:

tensor([ 1.0000e+00, -4.3711e-08, -1.0000e+00])

Copy to clipboard
Copy to clipboard

Example 2: Applying .cos() with a 2D tensor

In this example, .cos() is applied to a 2D tensor of angles in radians:

import torch

# Create a 2x2 tensor
matrix = torch.tensor([[0, torch.pi / 3], [torch.pi / 2, torch.pi]])

# Compute the cosine
result = torch.cos(matrix)

print(result)

Copy to clipboard
Copy to clipboard

The output of this code is:

tensor([[ 1.0000e+00,  5.0000e-01],
        [-4.3711e-08, -1.0000e+00]])

Copy to clipboard
Copy to clipboard