Write Conv2D and Active function from scratch

https://github.com/SamratSahoo/UT-Arlington-Research/blob/master/Week 6 - Convolutions %26 Wavelet Transforms/Convolutions.py

import numpy as np

def convolve2D(image, kernel, padding=0, strides=1):
    # Cross Correlation
    kernel = np.flipud(np.fliplr(kernel))

    # Gather Shapes of Kernel + Image + Padding
    xKernShape, yKernShape = kernel.shape
    xImgShape, yImgShape = image.shape

    # Shape of Output Convolution
    xOutput = (xImgShape - xKernShape + 2 * padding) // strides + 1
    yOutput = (yImgShape - yKernShape + 2 * padding) // strides + 1
    output = np.zeros((xOutput, yOutput))

    # Apply Padding
    if padding != 0:
        imagePadded = np.pad(image, padding, mode='constant')
    else:
        imagePadded = image

    # Iterate through image
    for x in range(xOutput):
        for y in range(yOutput):
            output[x, y] = np.sum(kernel * imagePadded[x * strides: x * strides + xKernShape,
                                                        y * strides: y * strides + yKernShape])
						
    return output