Machine Learning with Docker and Kubernetes: Training models

In this chapter, we will work on Kubernetes Jobs and how we can use these Jobs to train a machine learning model. A job creates one or more Pods. It is a Kubernetes controller making sure that the Pods successfully terminate their workload. The Job is considered complete when a specified number of Pods terminate. If a Pod fails, the Job will create a new Pod to replace it. Our goal will be to create a Kubernetes Job, let's call it "training Job", that loads training data stored in GitHub, train a machine learning model, serialize the model, and store it outside the cluster.