DQN-Atari-Tensorflow

Reimplementing "Human-Level Control Through Deep Reinforcement Learning" in Tensorflow

This may be the simplest implementation of DQN to play Atari Games.

The pretrained network would release soon!

Prerequsite

1. Tensorflow (prefer with GPU CUDA supported)
2. opencv2
3. Arcade Learning Environment ( https://github.com/mgbellemare/Arcade-Learning-Environment )

Setup and Run

git clone https://github.com/songrotek/DQN-Atari-Tensorflow.git

cd DQN-Atari-Tensorflow
python AtariDQN.py

Reference

1. asrivat1/DeepLearningVideoGames
2. gliese581gg/DQN_tensorflow