Obtaining the dataset
Once you have a task that you want to perform with a neural network, the first step is usually to obtain the dataset, which is the data that you need to feed to the neural network. In the tasks that we perform in this book, the dataset is usually composed of images or videos, but it could be anything, or a mix of images and other data.
The dataset represents the input that you feed to your neural network, but as you may have noticed, your dataset also contains the desired output, the labels. We will call x the input to the neural network, and y the output. The dataset is composed of the inputs/features (for example, the images in the MNIST dataset), and the output/labels (for example, the number associated with each image).
We have different dataset types. Let's start with the easiest – the datasets included in Keras – before proceeding to the next ones.
Datasets in the Keras module
Usually a dataset is a lot of data. It's normal...
