PyTorch .neg()

The .neg() method in PyTorch returns a new tensor with the negative of each element in the input tensor. This operation multiplies each element by -1, effectively flipping the sign of all values. The method is commonly used in mathematical operations, gradient computations, and transformations in neural networks.

Syntax

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

Parameters:

• input (Tensor): The input tensor.
• out (Tensor, optional): The output tensor to store the result. Must have the same shape as input.

Return value:

Returns a new tensor where each element is the negative of the corresponding element in input.

Example

The following example demonstrates how to use the .neg() method to negate tensor elements:

import torch

# Create a tensor with positive and negative values
tensor = torch.tensor([1.5, -2.3, 0.0, 4.8, -1.2])

# Compute the negative using the method form
neg_tensor = tensor.neg()

# Alternative: use the functional form
neg_functional = torch.neg(tensor)

# Alternative: use the operator form
neg_operator = -tensor

print("Original Tensor:")
print(tensor)

print("\nNegated Tensor (using .neg()):")
print(neg_tensor)

print("\nNegated Tensor (using torch.neg()):")
print(neg_functional)

print("\nNegated Tensor (using - operator):")
print(neg_operator)

Copy to clipboard
Copy to clipboard

This example results in the following output:

Original Tensor:
tensor([ 1.5000, -2.3000,  0.0000,  4.8000, -1.2000])

Negated Tensor (using .neg()):
tensor([-1.5000,  2.3000, -0.0000, -4.8000,  1.2000])

Negated Tensor (using torch.neg()):
tensor([-1.5000,  2.3000, -0.0000, -4.8000,  1.2000])

Negated Tensor (using - operator):
tensor([-1.5000,  2.3000, -0.0000, -4.8000,  1.2000])

Copy to clipboard
Copy to clipboard

In this example:

• Positive values become negative: 1.5 → -1.5, 4.8 → -4.8
• Negative values become positive: -2.3 → 2.3, -1.2 → 1.2
• Zero remains zero: 0.0 → -0.0 (negative zero in floating-point)
• All three forms (.neg(), torch.neg(), and -) produce identical results