We can use numpy nd-array to create a tensor in Python. We can use the following Python code to perform tensor addition and subtraction.

import numpy

A = numpy.random.randint(low=1, high=10, size=(3, 3, 3))
B = numpy.random.randint(low=1, high=10, size=(3, 3, 3))

C = A + B
D = A - B
print("A: \n", A)
print("B: \n", B)
print("C = A + B: \n", C)
print("D = A - B: \n", D)

Here, we are using randint() function from the numpy.random module to generate random integers within the range [low, high). And the shape of the created tensor is given by the size argument. For example, size = (3, 3, 3) indicates that the tensor has 3 elements across each dimension.

After that, we are adding the two tensors A and B to create the tensor C. And the tensor D contains the tensor (A – B). The output of the given program will be:

A: 
 [[[7 5 5]
  [4 9 3]
  [2 4 8]]

 [[3 4 2]
  [8 6 4]
  [4 1 2]]

 [[1 9 4]
  [3 3 6]
  [3 3 6]]]
B: 
 [[[7 9 5]
  [5 8 4]
  [2 6 6]]

 [[6 6 6]
  [1 4 7]
  [6 2 5]]

 [[2 9 1]
  [9 8 9]
  [4 4 1]]]
C = A + B: 
 [[[14 14 10]
  [ 9 17  7]
  [ 4 10 14]]

 [[ 9 10  8]
  [ 9 10 11]
  [10  3  7]]

 [[ 3 18  5]
  [12 11 15]
  [ 7  7  7]]]
D = A - B: 
 [[[ 0 -4  0]
  [-1  1 -1]
  [ 0 -2  2]]

 [[-3 -2 -4]
  [ 7  2 -3]
  [-2 -1 -3]]

 [[-1  0  3]
  [-6 -5 -3]
  [-1 -1  5]]]