Implementing an advanced CNN

It is crucial to be able to extend CNN models for image recognition so that we understand how to increase the depth of the network. This way, we may increase the accuracy of our predictions if we have enough data. Extending the depth of CNN networks is done in a standard fashion: we just repeat the convolution, max pooling, and ReLU in series until we are satisfied with the depth. Many of the more accurate image recognition networks operate in this fashion.

Loading and preprocessing data may cause a big headache: most image datasets will be too large to fit into memory, but image preprocessing will be needed to improve the performance of the model. What we can do with TensorFlow is use the tf.data API to create an input pipeline. This API contains a set of utilities for loading and preprocessing data. Using it, we will instantiate a tf.data.Dataset object from the CIFAR-10 dataset (downloaded through the Keras dataset API tf.keras.datasets...