One of the most commonly used operations in NumPy is the sum function, which allows you to calculate the sum of all the elements in an array or matrix. However, sometimes you may need to sum only over specific axes of a multi-dimensional array. In this article, we will discuss how to get the sum over the last two axes at once using NumPy.
1. Using “np.reshape()”, “np.sum()”, and “np.prod()” functions:
The code defines a function sum_last_two_axes that takes a NumPy array with four dimensions as input returns the sum of its last two axes. To achieve this, the function first calculates the product of all dimensions except the last two using np.prod(shape[:-2]). Then it reshapes the input array to collapse the last two dimensions and create a 2D array. Next, the function sums over the last axis of the reshaped array using reshaped_arr.sum(axis=-1) to get the sum over the last two axes. Finally, it reshapes the resulting array to have the same shape as the original input array, except for the last two dimensions which are collapsed. The code includes an example usage of the function.
2. Using the “np.transpose()” and “np.sum()” functions:
This code defines a function sum_last_two_axes that takes a multi-dimensional NumPy array “A” and returns the sum of its last two axes. First, the function transposes “A” so that the last two axes become the first two axes, then reshapes the array so that the last two axes are flattened into one axis. Next, it computes the sum along the flattened axis, and finally reshapes the sum array to match the original shape of the input array. The code includes an example usage of the function.
3. Using “np.transpose()” and “np.reshape()” functions:
This code defines a function sum_last_two_axes that takes a 4D NumPy array “A” and returns a new array that is the sum of the last two axes of “A”. The function transposes the last two axes of “A”, reshapes the resulting array so that the last two axes are flattened into one axis, and then sums along this flattened axis. The resulting array has the same shape as “A” except that the last two axes are collapsed into one. The code includes an example usage of the function.