RLCard Tutorial

This is an official tutorial for RLCard: A Toolkit for Reinforcement Learning in Card Games. We provide step-by-step instructions and running examples with Jupyter Notebook for both Python and R. The Python tutorial is available in Colab, where you can try your experiments in the cloud interactively.

• Official Website: http://www.rlcard.org
• Paper: https://arxiv.org/abs/1910.04376
• Resources: Awesome-Game-AI

For Python

Tutorials in Jupyter Notebook

• Training DQN on Blackjack
• Training CFR on Leduc Hold'em
• Having Fun with Pretrained Leduc Model
• Running multiple processes
• Leduc Hold'em as single-agent environment

Links to Colab

• Training DQN on Blackjack
• Training CFR on Leduc Hold'em
• Having Fun with Pretrained Leduc Model
• Running multiple processes
• Leduc Hold'em as single-agent environment

For R

This tutorial uses reticulate to call RLCard with R interfaces. Please make sure that you have Python 3.5+ and pip installed.

• Running Random agent on Blackjack
• Training DQN on Blackjack
• Training CFR on Leduc Hold'em
• Leduc Hold'em as single-agent environment
• Texas Hold'em nolimit

Contributing

Contribution to this project is greatly appreciated! Please create an issue/pull request for feedbacks or more tutorials.

Cite this work

If you find this repo useful, you may cite:

@article{zha2019rlcard,
  title={RLCard: A Toolkit for Reinforcement Learning in Card Games},
  author={Zha, Daochen and Lai, Kwei-Herng and Cao, Yuanpu and Huang, Songyi and Wei, Ruzhe and Guo, Junyu and Hu, Xia},
  journal={arXiv preprint arXiv:1910.04376},
  year={2019}
}

Acknowledgements

The R tutorial is mainly based on the code provided by @systats. See here.