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!
git clone https://github.com/songrotek/DQN-Atari-Tensorflow.git
cd DQN-Atari-Tensorflow
python AtariDQN.py