rlcard
Basic Information
This repository is focused on reinforcement learning and AI bots for card games. It collects implementations and resources geared toward developing, running, and studying RL agents in card-play domains. The project name and description list supported titles such as Blackjack, Leduc, Texas, DouDizhu, Mahjong, and UNO, indicating a breadth of poker-style and other card games. The repository is intended as a practical toolkit and reference for users who want to experiment with multi-agent interactions, imperfect information, and game-specific strategies in card environments. Its primary purpose is to support creation, evaluation, and comparison of RL-based card game agents.
Links
Stars
3215
Github Repository
Categorization
App Details
Features
Multi-game support covering several card titles including Blackjack, Leduc, Texas, DouDizhu, Mahjong, and UNO. Contains AI bot implementations and game environments to facilitate experiments with reinforcement learning in card settings. Organized to let users work across poker-style and non-poker card domains within a single project. Provides code artifacts that enable running agents against each other and reusing implementations for multiple games. The repository is oriented to practical agent development and cross-game experimentation rather than providing unrelated tooling.
Use Cases
The project helps researchers and developers prototype and assess reinforcement learning methods in a variety of card games, enabling study of multi-agent and imperfect-information scenarios. It is useful for educators demonstrating RL concepts in game settings and for practitioners who want ready examples of card-game agents. By offering multiple game environments and agent implementations, the repository supports comparative evaluation of algorithms, rapid experimentation with strategies, and reuse of game-specific code when building or testing new RL approaches.
-
$$$$ -
$$$$