Classic reinforcement learning, made inspectable

gym_classics2 is a teaching package containing finite Markov decision processes, textbook reinforcement-learning algorithms, and visualization tools. Its environments implement the standard Gymnasium API and also expose their transition models for planning algorithms.

Where to begin

• Follow Getting started to install the package and run an environment.
• Browse Environments to choose a task.
• Read Model access before using value or policy iteration.
• Use Choosing an algorithm to check an algorithm's requirements and outputs.
• Consult the API reference for exact signatures.

Design goals

The implementations favor correspondence with Sutton and Barto's pseudocode and inspectable intermediate results over framework abstractions. They are intended for experiments, demonstrations, and coursework rather than large-scale reinforcement-learning workloads.