RLCard: A Toolkit for Reinforcement Learning in Card Games

RLCard is a toolkit for Reinforcement Learning (RL) in card games. It supports multiple card environments with easy-to-use interfaces. The goal of RLCard is to bridge reinforcement learning and imperfect information games, and push forward the research of reinforcement learning in domains with multiple agents, large state and action space, and sparse reward. RLCard is developed by DATA Lab at Texas A&M University.

• Official Website: http://www.rlcard.org
• Paper: https://arxiv.org/abs/1910.04376

News:

• PyTorch implementation available. Thanks for the contribution of @mjudell.
• We have just initialized a list of Awesome-Game-AI resources. Check it out!

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}
}

Installation

Make sure that you have Python 3.5+ and pip installed. We recommend installing rlcard with pip as follow:

git clone https://github.com/datamllab/rlcard.git
cd rlcard
pip install -e .

or use PyPI with:

pip install rlcard

To use tensorflow implementation, run:

pip install rlcard[tensorflow]

To try out PyTorch implementation for DQN and NFSP, please run:

pip install rlcard[torch]

If you meet any problems when installing PyTorch with the command above, you may follow the instructions on PyTorch official website to manually install PyTorch.

Examples

Please refer to examples/. A short example is as below.

import rlcard
from rlcard.agents.random_agent import RandomAgent

env = rlcard.make('blackjack')
env.set_agents([RandomAgent(action_num=env.action_num)])

trajectories, payoffs = env.run()

We also recommend the following toy examples.

• Playing with random agents
• Deep-Q learning on Blackjack
• Running multiple processes
• Having fun with pretrained Leduc model
• Leduc Hold'em as single-agent environment
• Training CFR on Leduc Hold'em

Demo

With tensorflow installed, run examples/leduc_holdem_human.py to play with the pre-trained Leduc Hold'em model. Leduc Hold'em is a simplified version of Texas Hold'em. Rules can be found here.

>> Leduc Hold'em pre-trained model

>> Start a new game!
>> Agent 1 chooses raise

=============== Community Card ===============
┌─────────┐
│░░░░░░░░░│
│░░░░░░░░░│
│░░░░░░░░░│
│░░░░░░░░░│
│░░░░░░░░░│
│░░░░░░░░░│
│░░░░░░░░░│
└─────────┘
===============   Your Hand    ===============
┌─────────┐
│J        │
│         │
│         │
│    ♥    │
│         │
│         │
│        J│
└─────────┘
===============     Chips      ===============
Yours:   +
Agent 1: +++
=========== Actions You Can Choose ===========
0: call, 1: raise, 2: fold

>> You choose action (integer):

Documents

Please refer to the Documents for general introductions. API documents are available at our website.

Available Environments

We provide a complexity estimation for the games on several aspects. InfoSet Number: the number of information sets; Avg. InfoSet Size: the average number of states in a single information set; Action Size: the size of the action space. Name: the name that should be passed to rlcard.make to create the game environment.

Game	InfoSet Number	Avg. InfoSet Size	Action Size	Name	Status
Blackjack (wiki, baike)	10^3	10^1	10^0	blackjack	Available
Leduc Hold’em (paper)	10^2	10^2	10^0	leduc-holdem	Available
Limit Texas Hold'em (wiki, baike)	10^14	10^3	10^0	limit-holdem	Available
Dou Dizhu (wiki, baike)	10^53 ~ 10^83	10^23	10^4	doudizhu	Available
Simple Dou Dizhu (wiki, baike)	-	-		simple-doudizhu	Available
Mahjong (wiki, baike)	10^121	10^48	10^2	mahjong	Available
No-limit Texas Hold'em (wiki, baike)	10^162	10^3	10^4	no-limit-holdem	Available
UNO (wiki, baike)	10^163	10^10	10^1	uno	Available
Sheng Ji (wiki, baike)	10^157 ~ 10^165	10^61	10^11	-	Developing

Evaluation

The perfomance is measured by winning rates through tournaments. Example outputs are as follows:

Library Structure

The purposes of the main modules are listed as below:

• /examples: Examples of using RLCard.
• /docs: Documentation of RLCard.
• /tests: Testing scripts for RLCard.
• /rlcard/agents: Reinforcement learning algorithms and human agents.
• /rlcard/envs: Environment wrappers (state representation, action encoding etc.)
• /rlcard/games: Various game engines.
• /rlcard/models: Model zoo including pre-trained models and rule models.

API Cheat Sheet

• rlcard.make(env_id, config={}): Make an environment. env_id is a string of a environment; config is a dictionary specifying some environment configurations, which are as follows.
  • allow_step_back: Defualt False. True if allowing step_back function to traverse backward in the tree.
  • allow_raw_data: Default False. True if allowing raw data in the state.
  • single_agent_mode: Default False. True if using single agent mode, i.e., Gym style interface with other players as pretrained/rule models.
  • active_player: Defualt 0. If single_agent_mode is True, active_player will specify operating on which player in single agent mode.
  • record_action: Default False. If True, a field of action_record will be in the state to record the historical actions. This may be used for human-agent play.
• env.step(action, raw_action=False): Take one step in the environment. action can be raw action or integer; raw_action should be True if the action is raw action (string).
• env.step_back(): Available only when allow_step_back is `True. Take one step backward. This can be used for algorithms that operate on the game tree, such as CFR.
• env.init_game(): Initialize a game. Return the state and the first player ID.
• env.get_payoffs(): In the end of the game, return a list of payoffs for all the players.
• env.run(): Run a complete game and return trajectories and payoffs. The function can be used after the agents are set up.
• State Definition: State will always have observation state['obs'] and legal actions state['legal_actions']. If allow_raw_data is True, state will have raw observation state['raw_obs'] and raw legal actions state['raw_legal_actions'].

Contributing

Contribution to this project is greatly appreciated! Please create an issue for feedbacks/bugs. If you want to contribute codes, please refer to Contributing Guide.

Acknowledgements

We would like to thank JJ World Network Technology Co.,LTD for the generous support.