/dqn

deep-q-learning

Primary LanguagePython

dqn

Tensorflow implementation of the deep Q-learning based on the DeepMind Nature paper in 2015 using OpenAI Gym as the environment. Environments and various hyper parameters as well as the option to enable Double Q-learning can be set in the configuration file.

Dependencies

  1. Tensorflow
  2. OpenCV
  3. Gym

Usage

  1. Edit config.py
  2. Run python train.py to start training
  3. Run python play.py to load the latest checkpoint and play