10. Getting Deep RL to Work
A deep RL system consists of an agent interacting with an environment. Agents further consist of components which include a memory, policy, neural networks, and algorithmic functions. Individually, each of these components can be quite complex, and they also have to integrate and operate together to produce a working deep RL algorithm. As a result, a code base that implements multiple deep RL algorithms starts to enter the realm of large software systems with substantial amounts of code. This complexity leads to interdependencies and constraints and leaves more room for bugs to creep in. Consequently, the software can be fragile and difficult to get working.
This chapter is filled with some practical tips for debugging ...
Get Foundations of Deep Reinforcement Learning: Theory and Practice in Python now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
