Chapter 8: Extended fastai and Deployment Features

So far in this book, you have learned how to ingest and explore datasets using fastai, how to train fastai models with tabular, text, and image datasets, and how to deploy fastai models. Throughout the book so far, the emphasis has been on covering as much of the functionality of fastai as possible using the highest-level fastai API. In particular, we have emphasized using dataloaders objects as the basis for defining the datasets used to train the model. Up to this point in the book, we have taken the happy path whenever possible. To demonstrate how to accomplish tasks using fastai, we have chosen the most straightforward way possible.

In this chapter, we are going to take some steps off the happy path to explore additional features of fastai. You will learn how to track what is happening with your model more closely, how to control the training process, and generally how to take advantage of more of the capabilities that fastai...