/AlphaZero

A simple implementation of DeepMind's AlphaZero on OpenAi Gym CartPole

Primary LanguagePython

AlphaZero-CartPole

A simple implementation of DeepMind's AlphaZero on OpenAI's Cartpole gym. Very simple implementation, tree depth is only set to 1

Usage

You can train the model by executing the following command:

python alphazero.py

Work in progress. Environment is consider solved if an agent is able to achieve and maintain a episodic reward of about 500!!

Dependencies

  • PyTorch
  • OpenAI gym

References