Compute Element-wise Angle of Tensor in PyTorch

To compute the elementwise angle of the given input tensor, we apply torch.angle(). It takes an input tensor and returns a tensor with angle in radian computed element wise. To convert the angles into the degree we multiply the angle in radian by 180/np.pi. It supports both real and complex-valued tensors.

Syntax

torch.angle(input)

Steps

To compute the elementwise angle, you could follow the steps given below −

• Import the required library. In all the following examples, the required Python library is torch. Make sure you have already installed it.

import torch

• Define torch tensors and print them.

input = torch.tensor([1 + 1j, -1 -4j, 3-2j])

• Compute torch.angle(input). It is a tensor with angle in radians computed elementwise for the input.

angle = torch.angle(input)

• Print the above computed tensor with angles in radians.

print("Angles in Radian:
", angle)

Let's take a couple of examples to demonstrate how to compute elementwise angles in radians and degrees.

Example 1

In the Python3 program below we compute element wise angle in radians and degrees of the given input tensor.

# Python 3 program to compute the element-wise
# angle (in radians and degree) of the given input tensor
# Import the required libraries
import torch
from numpy import pi

# define a complex input tensor
input = torch.tensor([1 + 1j, -1 -4j, 3-2j])

# print the input tensor
print("Input Tensor:
", input)

# compute the angle in radians
angle = torch.angle(input)

# print the computed tensor of angles
print("Angles in Radian:
", angle)

# convert the angle s in degree
degree = angle*180/pi

# print the computed tensor of degree
print("Angles in Degree:
", degree)

Output

Input Tensor:
   tensor([ 1.+1.j, -1.-4.j, 3.-2.j])
Angles in Radian:
   tensor([ 0.7854, -1.8158, -0.5880])
Angles in Degree:
   tensor([ 45.0000, -104.0362, -33.6901])

Example 2

In this program, we compute element wise angle in radians and degrees of the given input tensor.

# Python 3 program to to compute the element-wise
# angle (in radians and degrees) of the given input tensor
# Import the required libraries
import torch
from numpy import pi

# define a complex input tensor
real = torch.randn(3,4)
imag = torch.randn(3,4)
input = torch.complex(real, imag)

# print the input tensor
print("Input Tensor:
", input)

# compute the angle in radians
angle = torch.angle(input)

# print the computed tensor of angles
print("Angles in Radian:
", angle)

# convert the angle s in degree
degree = angle*180/pi

# print the computed tensor of degree
print("Angles in Degree:
", degree)

Output

Input Tensor:
   tensor([[ 0.1967-0.0188j, -0.5311-1.2427j, 0.9937+0.0051j, - 1.8304-0.1321j],
      [-0.1787+1.7834j, 0.9925+0.2452j, -0.8813-0.0207j, - 0.4967+0.9938j],
      [-0.9051-0.1204j, 1.0013+0.3430j, 0.6131-0.0317j, - 0.3861+0.6365j]])
Angles in Radian:
   tensor([[-0.0955, -1.9747, 0.0052, -3.0695],
      [ 1.6707, 0.2422, -3.1181, 2.0343],
      [-3.0093, 0.3300, -0.0517, 2.1161]])
Angles in Degree:
   tensor([[ -5.4711, -113.1396, 0.2962, -175.8722],
      [ 95.7231, 13.8750, -178.6560, 116.5553],
      [-172.4209, 18.9103, -2.9624, 121.2437]])