PyTorch .lgamma()

In PyTorch, the .lgamma() function returns the natural logarithm of the absolute value of the gamma function for each element in an input tensor.

Syntax

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

Parameters:

• input (Tensor): Input tensor of floating point or complex numbers.
• out (Optional): Output tensor to write results to.

Return value:

Returns a tensor of the same shape as the input, where each element is the log of the absolute gamma function of the corresponding input element.

Example

The example demonstrates computing the natural logarithm of the absolute gamma function using both the functional form torch.lgamma() and the tensor method .lgamma(), then verifies both approaches produce identical results:

import torch

x = torch.tensor([0.5, 1.0, 2.5, 5.0, -0.5])

# Functional form
y1 = torch.lgamma(x)

# Tensor method
y2 = x.lgamma()

print(y1)
print(torch.allclose(y1, y2))

Copy to clipboard
Copy to clipboard

Here’s the output of this code:

tensor([0.5724, 0.0000, 0.2847, 3.1781, 1.2655])
True

Copy to clipboard
Copy to clipboard