muzero-general and muzero

The kaesve implementation is a specialized fork/reimplementation of the werner-duvaud framework that adds MuZero-specific features (learned MDP models, inter-algorithm comparison) while maintaining compatibility with the AlphaZero General architecture, making them ecosystem variants rather than true competitors or complements.

muzero-general
51
Established
muzero
44
Emerging
Maintenance 0/25
Adoption 10/25
Maturity 16/25
Community 25/25
Maintenance 0/25
Adoption 10/25
Maturity 16/25
Community 18/25
Stars: 2,784
Forks: 670
Downloads:
Commits (30d): 0
Language: Python
License: MIT
Stars: 168
Forks: 27
Downloads:
Commits (30d): 0
Language: Jupyter Notebook
License: MIT
Stale 6m No Package No Dependents
Stale 6m No Package No Dependents

About muzero-general

werner-duvaud/muzero-general

MuZero

Implements the DeepMind MuZero algorithm using PyTorch with residual and fully-connected networks, enabling model-based RL without environment dynamics knowledge. Supports distributed training across multiple GPUs and Ray clusters with asynchronous self-play, plus real-time TensorBoard monitoring. Designed for quick adaptation to new environments—board games, Atari, and OpenAI Gym—by simply defining game classes and hyperparameters.

About muzero

kaesve/muzero

A clean implementation of MuZero and AlphaZero following the AlphaZero General framework. Train and Pit both algorithms against each other, and investigate reliability of learned MuZero MDP models.

Related comparisons

Scores updated daily from GitHub, PyPI, and npm data. How scores work