A simplified, highly flexible, commented and (hopefully) easy to understand implementation of self-play based reinforcement learning based on the AlphaGo Zero paper (Silver et al). It is designed to be easy to adopt for any two-player turn-based adversarial game and any deep learning framework of your choice. A sample implementation has been provided for the game of Othello in PyTorch, Keras and TensorFlow. An accompanying tutorial can be found here. We also have implementations for GoBang and TicTacToe.
To use a game of your choice, subclass the classes in Game.py and NeuralNet.py and implement their functions. Example implementations for Go can be found in go/GoGame.py and go/{pytorch,keras,tensorflow}/NNet.py.
GoCoach.py contains the core training loop and GoMCTS.py performs the Monte Carlo Tree Search. The parameters for the self-play can be specified in main.py. Additional neural network parameters are in go/{pytorch,keras,tensorflow}/NNet.py (cuda flag, batch size, epochs, learning rate etc.).
To start training a model for GoGame:
python main.pyChoose your framework and game in main.py.
To display the board status of training process: set display:True in main.py file.
To start pitting with a player for GoGame:
python pit.pyWe trained a PyTorch model for 6x6 Othello (~80 iterations, 100 episodes per iteration and 25 MCTS simulations per turn). This took about 3 days on an NVIDIA Tesla K80. The pretrained model (PyTorch) can be found in pretrained_models/othello/pytorch/. You can play a game against it using pit.py. Below is the performance of the model against a random and a greedy baseline with the number of iterations.
A concise description of our algorithm can be found here.
While the current code is fairly functional, we could benefit from the following contributions:
- Game logic files for more games that follow the specifications in
Game.py, along with their neural networks - Neural networks in other frameworks
- Pre-trained models for different game configurations
- An asynchronous version of the code- parallel processes for self-play, neural net training and model comparison.
- Asynchronous MCTS as described in the paper
- Shantanu Thakoor and Megha Jhunjhunwala helped with core design and implementation.
- Shantanu Kumar contributed TensorFlow and Keras models for Othello.
- Evgeny Tyurin contributed rules and a trained model for TicTacToe.
- MBoss contributed rules and a model for GoBang.
- TianyuLei & LingxiaoGong contributed rules and a model for GoGame.
- ZhongyaoChu contributed ResNet implementation and improved MCTS for GoGame.
Thanks to pytorch-classification and progress.
- Pytorch with CUDA support is required
- If a required environment is not available, please contact zc1213@nyu.edu and request for a link on which online pitting experiecne is offered.