Compute Inverse Cosine and Inverse Hyperbolic Cosine in PyTorch

The torch.acos() method computes the inverse cosine of each element of an input tensor. It supports both real and complex-valued inputs. It supports any dimension of the input tensor. The elements of the input tensor must be in the range [-1,1], as the inverse cosine function has its domain as [-1,1].

The torch.acosh() method computes the inverse hyperbolic cosine of each element of the input tensor. It also supports both real and complex-valued inputs of any dimension. The elements of the input tensor must be any number greater or equal to 1, as the inverse cosine function has its domain as [1, +∞].

Syntax

torch.acos(input)
torch.acosh(input)

Steps

To compute the inverse cosine and inverse hyperbolic cosine of each element in the input tensor, you could follow the below steps −

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

import torch

• Create a torch tensor and print it. To compute the inverse cosine we ensure that input tensor elements are in range [-1,1] by applying uniform_(-1,1) to generate random numbers.

a = torch.randn(4).uniform_(-1, 1)
print("Input Tensor:
", input)

• Compute the inverse cosine or inverse hyperbolic cosine of each element in the input tensor using torch.acos(input) or torch.acosh(input). Here input is the input tensor.

inv_cos = torch.acos(input)
inv_cosh = torch.acosh(input)

• Display the computed tensors with inverse cosine or inverse hyperbolic cosine values.

print("Inverse Cosine Tensor:
", inv_cos)
print("Inverse Hyperbolic Cosine Tensor:
", inv_cosh)

Example 1

import torch

# define a tensor with values in range [-1, 1]
a = torch.randn(4).uniform_(-1, 1)

# print the above defined tensors
print("Tensor:",a)

# compute inverse cosine of elements of the above tensor
inv_cos = torch.acos(a)

# print the tensor with inverse cosine values
print("Inverse Cosine:", inv_cos)

Output

Tensor: tensor([ 0.2127, 0.8572, -0.3944, -0.9310])
Inverse Cosine: tensor([1.3565, 0.5409, 1.9762, 2.7679])

Example 2

import torch

# define a tensor with values in range [-1, 1]
a = torch.randn(5,5).uniform_(-1, 1)

# print the above defined tensors
print("Tensor:
",a)

# compute inverse cosine of elements of the above tensor
inv_cos = torch.acos(a)

# print the tensor with inverse cosine values
print("Inverse Cosine:
", inv_cos)

Output

Tensor:
   tensor([[ 0.6149, -0.6334, 0.8994, 0.6377, -0.2348],
      [-0.1458, -0.2893, 0.7044, -0.9379, 0.3848],
      [-0.9450, 0.0991, -0.8826, 0.8640, 0.7513],
      [ 0.0019, -0.2069, 0.6228, 0.8062, -0.9137],
      [-0.8181, 0.4544, -0.8216, -0.7370, 0.9821]])
Inverse Cosine:
   tensor([[0.9085, 2.2568, 0.4525, 0.8792, 1.8078],
      [1.7171, 1.8643, 0.7892, 2.7874, 1.1758],
      [2.8085, 1.4715, 2.6521, 0.5277, 0.7207],
      [1.5689, 1.7792, 0.8984, 0.6331, 2.7230],
      [2.5288, 1.0991, 2.5350, 2.3995, 0.1895]])

Example 3

import torch

# define a tensor with values in range [1, 9]
# the upper limit may be any number
a = torch.randn(4).uniform_(1, 9)
print("Tensor:",a)

# compute inverse hyperbolic cosine of above tensor
inv_cosh = torch.acosh(a)

# print the tensor with inverse hyperbolic cosine values
print("Inverse Hyperbolic Cosine:", inv_cosh)

Output

Tensor: tensor([4.7254, 8.4879, 7.2126, 3.6867])
Inverse Hyperbolic Cosine: tensor([2.2347, 2.8283, 2.6641, 1.9790])

Example 4

import torch

# define a tensor with values in range [1, 9]
# the upper limit may be any number
a = torch.randn(4,4).uniform_(1, 9)
print("Tensor:
",a)

# compute inverse hyperbolic cosine of above tensor
inv_cosh = torch.acosh(a)

# print the tensor with inverse hyperbolic cosine values
print("Inverse Hyperbolic Cosine:
", inv_cosh)

Output

Tensor:
   tensor([[3.2647, 4.1873, 3.1671, 2.4577],
      [4.4552, 8.1373, 1.2274, 5.5870],
      [2.6106, 1.5915, 2.5918, 7.5412],
      [7.6130, 8.9626, 4.6831, 1.1207]])
Inverse Hyperbolic Cosine:
   tensor([[1.8520, 2.1106, 1.8200, 1.5481],
      [2.1744, 2.7858, 0.6623, 2.4055],
      [1.6138, 1.0402, 1.6060, 2.7091],
      [2.7187, 2.8831, 2.2255, 0.4865]])