Observe your machine learning to play a video game

Description

This application of Reinforcement Learning (Markov Decision Process) is using features like Dueling-DQN, POMDP, and experience replay to allow you to watch your computer learning to play the CartPole-v0 game in the Gym library.

You can choose how many sessions you want your computer to play with the [-n] option. By default it will play 5000 games. example :

python Codes/Training.py -n 1000

Installation

You can directly clone the repository and install the packages listed in the requirements.txt file. Here are the commands with python in 3.6.5 :

git clone https://github.com/VBrabant/Observe_ReinforcementLearning
pip install -r requirements.txt