README.md

tetrisRL

demo video

Results

play on 7x14 board

Quick start

Installation

1. Clone this repo

  $ git clone https://github.com/SayhoKim/tetrisRL.git

2. Install Pygame, Keras and Tensorflow

  $ cd {Project path}
  $ pip3 install -r requirements.txt

Training

  $ python3 main.py

Demo

After editing configuration file (MODELPATH):

  $ python3 main.py --demo

Reference

1. Playing Atari with Deep Reinforcement Learning[arxiv]
2. Dueling Network Architectures for Deep Reinforcement Learning[arxiv]
3. Deep Reinforcement Learning with Double Q-learning[arxiv]
4. Prioritized Experience Replay[arxiv]