In the previous chapter, you learned how to train and deploy models with built-in frameworks such as scikit-learn or TensorFlow. Thanks to Script Mode, these frameworks make it easy to use your own code, without having to manage any training or deployment containers.
In some cases, your business or technical environment could make it difficult or even impossible to use these containers. Maybe you need to be in full control of how containers are built. Maybe you'd like to implement your own prediction logic. Maybe you're working with a framework or a language that's not natively supported by SageMaker.
In this chapter, you'll learn how to tailor training and prediction containers to your own needs. You'll also learn how to train and deploy your own custom code, using either the SageMaker SDK directly or command-line open source tools. We'll cover the following topics:
- Understanding how SageMaker invokes...