pokemon-deep-learning

An openAI gym project for pokemon involving deep q learning. Made by myself, Sam Little, and Layton Webber.

This code captures games played online, interprets them, updates a RNN to learn from them, and implements and evaluates them against a random agent.

Requires a locally hosted node.js Pokemon Showdown server and poke-env.

Gym can be installed with 'pip install -e agent_vs_agent'.

Program can be run with 'python3 main.py'.