Skip to content

Instantly share code, notes, and snippets.

@andreimuntean

andreimuntean/README.md

Created March 11, 2017 19:45
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save andreimuntean/0b695674b88e5181c30693982226e451 to your computer and use it in GitHub Desktop.
Save andreimuntean/0b695674b88e5181c30693982226e451 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
README.md

Deep Q-Learning

Deep reinforcement learning for environments with small state spaces.

The CartPole Gym environment was mastered in just a few minutes of training. Training and evaluation code is available at github.com/andreimuntean/deep-q-learning-lite.

Dependencies

  • OpenAI Gym 0.8
  • TensorFlow 1.0

Learning Environment

Uses environments provided by OpenAI Gym.

Network Architecture

The network has a single hidden layer with 40 rectified linear units. The output layer has as many nodes as there are actions. Each output node represents the expected utility of an action.

Acknowledgements

Heavily influenced by DeepMind's seminal paper 'Playing Atari with Deep Reinforcement Learning' (Mnih et al., 2013) and 'Human-level control through deep reinforcement learning' (Mnih et al., 2015).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment