jonathan-laurent/AlphaZero.jl

Health v2 · maintenance only

Adoption not part of the score

1333 stars · 146 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A generic, simple, and fast Julia implementation of DeepMind's AlphaZero algorithm for training game-playing agents via self-play and MCTS. It is designed to be accessible to students and researchers while being fast enough to solve nontrivial games on a desktop computer with a GPU.

Use cases

• train an AlphaZero agent for Connect Four
• implement AlphaZero for a custom board game
• run reinforcement learning self-play experiments on a single GPU
• research combining learning and tree search
• reproduce AlphaZero results without a large compute cluster
• learn how the AlphaZero algorithm works internally

When to choose

• you want a hackable, readable AlphaZero implementation in pure Julia
• you need to train agents for nontrivial games on limited hardware
• you want generic interfaces to plug in new games or learning frameworks
• you need distributed training across a cluster without code changes

When to avoid

• you need a production-strength engine optimized for distributed C++/GPU clusters like Leela Zero
• your project is Python-based and you need JAX or PyTorch integration
• you are working outside two-player perfect-information games

Facets

library · maturity active

machine-learning deep-learning reinforcement-learning gpu-computing reinforcement-learning machine-learning deep-learning windows cross-platform alphazero mcts self-play game-ai reinforcement-learning game-development julia gpu linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="jonathan-laurent/AlphaZero.jl")