An implementation of the deep Q-network for Atari games of https://doi.org/10.1038/nature14236.

• Python 3
• TensorFlow 1.0
• NumPy
• OpenAI Gym
• Pillow

Please read https://chan-y-park.github.io/blog/rl_atari.html for a detailed description of this project.

This is the animation of playing Breakout after 5 million weight updates.