Kaixhin/Rainbow
A PyTorch implementation of Rainbow DQN combining seven deep reinforcement learning improvements for Atari game agents.
Not currently ranked — collecting fresh signals.
star history
This repository implements the Rainbow algorithm, which integrates seven key improvements to Deep Q-Networks: Double DQN, Prioritised Experience Replay, Dueling Networks, Multi-step Returns, Distributional RL, and Noisy Nets. It uses PyTorch as the deep learning framework and is designed to train reinforcement learning agents on Atari games. The project includes both the original Rainbow configuration and a data-efficient variant.
Frequently asked
- What is Kaixhin/Rainbow?
- A PyTorch implementation of Rainbow DQN combining seven deep reinforcement learning improvements for Atari game agents.
- Is Rainbow open source?
- Yes — Kaixhin/Rainbow is open source, released under the MIT license.
- What language is Rainbow written in?
- Kaixhin/Rainbow is primarily written in Python.
- How popular is Rainbow?
- Kaixhin/Rainbow has 1.7k stars on GitHub.
- Where can I find Rainbow?
- Kaixhin/Rainbow is on GitHub at https://github.com/Kaixhin/Rainbow.