Useful Tools and Tips

In this section, you will first learn the importance of different splits of the dataset. After that, you learn some tips that will come handy when you start working on datasets that haven't been processed before. Then come tools such as pandas profiling and TensorBoard, which will make your life easier by providing easy access to information. We will take a look at AutoML and how it can be used to get high-performance models without much manual effort. Finally, we will visualize our Keras model and export the model diagram to a file.

Train, Development, and Test Datasets

We briefly talked about train, development, and test datasets in the previous chapters. Here, we will delve deeper into the topic.

The training, or train set is a sample from the dataset, and we use this to create our machine learning models. The development, or dev set (also known as validation set), is a sample that helps us tune the hyperparameters of the created model. The testing or test set...