Learning Ray

Chapter 11. Ray’s Ecosystem and Beyond

Over the course of this book, you’ve seen many examples of Ray’s ecosystem. Now it’s time to take a more systematic approach and show you the full extent of integrations currently available for Ray. We do so by discussing this ecosystem as seen from Ray AIR so that we can discuss it in the context of a representative AIR workflow.

Clearly, we can’t give you concrete code examples for a majority of the libraries in Ray’s ecosystem. Instead, we have to be content with giving you another Ray AIR example showcasing some integrations and discussing what others are available and how to use them. Where appropriate, we’ll point you to more advanced resources to deepen your understanding.

Now that you know much more about Ray and its libraries, this chapter is also the right place to compare what Ray offers to similar systems. As you’ve seen, Ray’s ecosystem is quite complex, can be seen from different angles, and is used for different purposes. That means many aspects of Ray can be compared to other tools in the market.

We’ll also comment on how to integrate Ray into more complex workflows in existing ML platforms. To wrap things up, we’ll give you an idea how to continue your journey of learning Ray after finishing this book.

Note

The notebook for this chapter is available on GitHub.

A Growing Ecosystem

To give you a glimpse of Ray’s ecosystem by means of a concrete example,¹ we’re going to show you how to use Ray AIR with data and models from ...