O

openai-gym

This project merges two of my past project a video game 3d and a ppo deep reinforcement learning algorithm. The objective of the project is:

create an agent capable of playing to the game correctly. allow the game to have an auto-play mode to show the commands and functionality of the user. create a new openAI gym environment that allow also communication with js application. Create an abstract-based class for the ppo, capable of be used for any game just changing state space, observations ad actions.

progetto per la valutazione dell'esame di reinforcement learning. Implementazione ppo evironment gym.