novastar111/pacman_v0_remap

该数据集包含自定义VisGym Pacman2D环境的行为克隆轨迹。每条记录代表一个成功的oracle episode，历史条目包括image_prev、image和image_next，图像是基于greyblue9/pacman-python视觉资源渲染的JPEG base64字符串。环境设置为7x7到12x12的网格，包含Pacman、食物、墙壁和一个确定性幽灵。幽灵通过BFS距离追逐Pacman，动作空间为四个方向。数据集分为easy和hard两个难度级别，每个级别包含训练和测试集，初始状态通过init_state_hash确保不重复。

This dataset contains behavior-cloning trajectories for the custom VisGym Pacman2D environment. Each row is one successful oracle episode. The history entries include `image_prev`, `image`, and `image_next`; no synthetic thinking traces are stored. Images are JPEG base64 strings rendered from the greyblue9/pacman-python visual assets used by the environment. Environment summary: Grid size is constrained to 7x7 through 12x12. Entities are Pacman, sparse food, walls, and exactly one deterministic ghost. The ghost chases Pacman by BFS distance with fixed tie-break order up, down, left, right. The movement action space is four directions. Splits include easy and hard difficulty levels, each with train and test sets, with initial states checked to be disjoint by `init_state_hash`.