To illustrate how a CNN works in practice, we will be building a model to recognize whether an object in a photo is a cat or not. The dataset we are using has more depth than this, but it would take a rather long time to train it to correctly classify everything. It is fairly trivial to extend the example to classify everything, but we would rather not be sitting there for a week waiting for the model to train.

For our example, we will be using the following structure: