Keras Deep Learning Cookbook

How to do it…

Let's create a simple CNN for the MNIST dataset that explains how to use all of the aspects of a CNN implementation, including the convolutional layers, pooling layers, and dropout layers. CNNs reduce the dimensions of the layers as we go deeper and increase the number of feature maps to detect more features and decrease the computational cost:

Let's import the required APIs:

import numpyfrom keras import backend as Kfrom keras.utils import np_utilsfrom keras.layers import Dense, Flatten, Dropoutfrom keras.layers.convolutional import Conv2D, MaxPooling2Dfrom keras.models import Sequentialimport matplotlib.pyplot as pltfrom