/deeprl

Universal library for deep reinforcement learning.

Primary LanguagePython

Deep Reinforcement Learning

Quick Start

TODO

Available algorithms

Algorithm Repository Example json Documentation
Discrete DeepQ deeprl TODO TODO
Asynchronous DeepQ (soon) deeprl TODO TODO

Available Simulators

Algorithm Repository Example json Documentation
Karpathy Game SimulatorZoo TODO TODO

Requirements

  • tensorflow == 0.7 for Automated Differentiation.
  • SimulatorZoo for the models

Documentation

  1. Quick start guide.
  2. Writing your own simulator.
  3. Adopting Existing Algorithm to different type of input.
  4. Contributing to this repository.

Contributors

  • stas-sl (found critical bug in Discrete Deepq)

Mentions