Kubeflow for Machine Learning

Chapter 2. Hello Kubeflow

Welcome to your first steps into the exciting world of Kubeflow!

First off, we’ll set up Kubeflow on your machine, or on a cloud provider. Then we’ll dive into a comprehensive example. The goal of this example is to get a model trained and start serving as quickly as possible. In some parts of the first section, it may seem like we are instructing you to mindlessly enter commands. While we want you to follow along, we strongly encourage you to revisit this chapter after you’ve finished the book to reflect on the commands you entered, and consider how much your understanding has grown while reading.

We’ll provide instructions for setting up and testing our example on a local machine and a link to instructions for performing the same on real clusters. While we will point you to the config files and OCI containers that are driving all of this, they are not the focus of this chapter; they will be covered in detail in subsequent chapters. The focus of this chapter is an end-to-end example that you can follow along with at home.

In future chapters we will dig into the “why” of everything we’re doing, we promise.

For now, just enjoy the ride.

Getting Set Up with Kubeflow

One of the great things about Kubeflow being built with Kubernetes is the ability to do our initial development and exploration locally, moving into more powerful and distributed tools later on. Your same pipeline can be developed locally and moved into a cluster.

Tip

Though you could get started ...