Variable-size, six-player Chinese Checkers environment

该数据集是一个为跳棋游戏定制化的多代理强化学习环境，采用PettingZoo平台开发，支持包括连跳在内的传统游戏规则。该环境能够对不同的多代理设置进行训练和评估，包括完全独立、共享编码器以及完全共享架构等。其规模支持6名玩家，并提供可变大小的棋盘。任务聚焦于多代理强化学习。

This dataset is a multi-agent reinforcement learning (MARL) environment customized for the game of checkers, developed using the PettingZoo platform. It supports traditional game rules including consecutive jumps, and enables training and evaluation across various multi-agent setups, such as fully independent, shared encoder, and fully shared architectures. The environment supports up to six players and offers variable-sized checkerboards, with the core task focusing on multi-agent reinforcement learning.