Ray Tutorials

The second-generation tutorials for the Ray Project (GitHub).

Setup

Prerequisite Software

The following must be installed on your machine:

Python 3.8
Pip (recent version)

We recommend using Anaconda, especially if you do lots of Python development and you need to define different environments for different projects.

Then run the following Pip command to install the other dependencies, including Ray:

pip install -r environment.yml

Tutorial Modules

There are many modules in this project, organized in subdirectories. Lessons are provided as Jupyter Lab notebooks. A number is embedded in the name to indicate the sequence you should follow when going through the lessons. In the cases where some notebooks can be studied in an arbitrary order, they are indicated with a letter, such as A. For example, algorithm04A.ipynb and algorithm04B.ipynb would indicate that you can go through either or both of these notebooks in any order, after going through the first through third lessons.

Here are the modules.

Ray Core

Directory: ray-core

This is the place to start. This module introduces you to Ray, why it was created, what problems it solves, how to use it, and how it works behind the scenes.

Ray RLlib

Directory: ray-rllib

Ray RLlib is Ray's system for reinforcement learning. This module begins with a "crash course" in RL concepts. It then explores several of the commonly-used algorithms and approaches for different applications.

Ray Tune and Ray SGD

Directory: ray-tune

Ray Tune is Ray's system for hyperparameter tuning. This module starts with an explanation of what hyperparameter tuning is for and the performances challenges doing it for many applications. Then the module explores how to use Tune, how it integrates with several popular ML frameworks, and the algorithms supported in Tune. The new Ray SGD module is also covered.

Ray Serve

Directory: ray-serve

Ray Serve is Ray's system for scalable model serving, with capabilities that also make it suitable for other web server applications. This module starts with an explanation of what's required in model serving, followed by a tour of the API with examples.

Troubleshooting

When you first start Ray in the very first lesson (i.e., use ray.init()), you may run into a few issues:

If you get an error like ... INFO services.py:... -- Failed to connect to the redis server, retrying., it probably means you are running a VPN on your machine. At this time, you can't use ray.init() with a VPN running. You'll have to stop your VPN for now.
If ray.init() worked (for example, you see a message like View the Ray dashboard at localhost:8265) and you're using a Mac, you may get several annoying dialogs asking you if you want to allow incoming connections for Python and/or Redis. Click "Accept" for each one and they shouldn't appear again during this tutorial. MacOS is trying to verify if these executables have been properly signed. Ray uses Redis. If you installed Python using Anaconda or other mechanism, then it probably isn't properly signed from the point of view of MacOS. To permanently fix this problem, see this StackExchange post.

deanwampler/ray-tutorials