Reinforcement Learning for Abalone

In this project reinforcement learning methods (namely Rainbow) are implemented for the game of Abalone.

It is very much based on the OpenAI Gym environment gym-abalone and on the Rainbow is all you need tutorial/implementation.

Setup

Like always, it is preferred to use an virtual environment.

To install gym-abalone, run:

git clone git@github.com:towzeur/gym-abalone.git
cd gym-abalone
pip install -e .

TODO: Requirements for Rainbow

To test a trained rainbow agent, modify the AGENT_FILE_PATH variable accordingly and run python test_rainbow-agent.py