Hands-On Reinforcement Learning for Games

Exploring RL Lib

RL Lib is based on the Ray project, which is essentially a Python job-based system. RL Lib is more like ML-Agents, where it exposes functionality using config files although, in the case of ML-Agents, the structure is completely run on their platform. Ray is very powerful but requires a detailed understanding of the configuration parameters and setup. As such, the exercise we show here is just to demonstrate the power and flexibility of Ray but you are directed to the full online documentation for further discovery on your own.

Open your browser to colab.research.google.com and follow the next exercise:

1. The great thing about using Colab is it can be quite easy to run and set up. Create a new Python 3 notebook and enter ...