OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Alpha Zero General

Alpha Zero General is an open-source implementation of AlphaZero-style self-play reinforcement learning for two-player, turn-based adversarial games. It provides a training loop, Monte Carlo Tree Search, and game and neural-network interfaces, with Othello examples in PyTorch and Keras.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Alpha Zero General — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you want a codebase to inspect or adapt for self-play training in a two-player, turn-based game, especially when Othello or the PyTorch and Keras examples are a useful starting point.

Who it is for

  • Developers adapting self-play training to a two-player, turn-based game.
  • Researchers and learners exploring game-playing models, Monte Carlo Tree Search, or reinforcement-learning experiments.
  • PyTorch or Keras users who want an Othello example to examine or extend.

Documented uses

  • Train a game-playing model through self-play using the included Othello example.
  • Adapt the code to another game by implementing the game and neural-network interfaces.
  • Adjust self-play and neural-network training parameters for experiments.
  • Study or reuse the training loop and MCTS implementation in game-playing research or education.
Before you choose

Requirements & limits

The repository is public and specifies the MIT license. Its README documents a Docker-based Jupyter setup using NVIDIA Docker, with PyTorch as the default container.

  • The included framework examples are PyTorch and Keras; using another framework involves supplying a compatible neural network implementation.
  • A new game needs game logic and a neural network that follow the repository's interfaces.
  • The README documents self-play parameters in main.py and additional neural-network parameters in the Othello implementations.
  • The Docker setup requires NVIDIA Docker, while the README presents it as an environment option.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start Othello training in the Docker container

Start training an Othello model using the repository’s Docker setup.

Before you start

  • nvidia-docker is set up.
  • The ./setup_env.sh setup step has created the default PyTorch Jupyter Docker container.
  • The repository files are available where the documented commands are run.

Steps

  1. In a terminal, run ./setup_env.sh to set up the default PyTorch Jupyter Docker container.

  2. Open a new terminal.

  3. Run the command below.

    Documented command
    docker exec -ti pytorch_notebook python main.py

Expected outcome

The documented command starts the Othello training entry point in the PyTorch container.

Guide limitations

  • The instructions describe choosing the framework and game in main.py; the Docker command uses the default PyTorch container.
  • No training duration or successful completion is guaranteed.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
suragnair/alpha-zero-general · Official homepage title · retained in its source language
How we review sourcesSuggest a correction