How to Compute the Error Function of a Tensor in PyTorch

Namachila October 20, 2025

0 0 1 minute read

In this article, we are going to cover how to compute the error function of a tensor in Python using PyTorch.

torch.special.erf() method

We can compute the error function of a tensor by using torch.special.erf() method. This method accepts the input tensor of any dimension and it returns a tensor with a computed error function with the same dimension as the input tensor. The below syntax is used to compute the error function of a tensor.

Syntax: torch.special.erf(input)

Parameters:

input: This is our input tensor.

Return: This method returns a tensor with computed error function of input tensor.

Example 1:

The following program is to understand how to compute the error function of the 1D tensor.

Python3

# import required libraries
import torch
 
# creating a 1D tensor
tens = torch.tensor([-0.7336, -0.9200, -0.4742,
                     -0.4470, -0.3472])
 
# print above created tensor
print("\n Input Tensor:", tens)
 
# compute the error function
er = torch.special.erf(tens)
 
# Display result
print("\n After Computed Error function :", er)

Output:

Example 2:

The following program is to know how to compute the error function of a batch of tensors.

Python3

# import required libraries
import torch
 
# creating a batch of tensor
tens = torch.tensor([[[0.8636, -0.4195, -0.4681],
                      [0.1265, 1.2233, 0.1978], 
                      [1.1389, 0.3686, 1.2339]],
                     [[1.6362, 0.6235, 1.2631], 
                      [0.3336, 1.5336, 1.3677], 
                      [0.5637, 1.3236, 0.2696]]])
 
# print above created tensor
print("\n\n Input Tensor: \n", tens)
 
# compute the error function
er = torch.special.erf(tens)
 
# Display result
print("\n\n After Computed Error function: \n", er)