werner-duvaud/muzero-general
A commented implementation of the MuZero deep reinforcement learning algorithm using PyTorch and Monte Carlo Tree Search.

Velocity · 7d
+1.2
★ / day
Trend
→steady
star history
This repository provides a documented implementation of MuZero, a Google DeepMind reinforcement learning algorithm that learns an internal model of the environment for planning. It uses PyTorch for neural network training, Monte Carlo Tree Search for decision making, and residual networks for function approximation. The project is designed for adaptability across different games and RL environments like OpenAI Gym.