lukedennison/chess-preprocessed

该数据集包含一系列谜题或游戏相关数据，可能涉及国际象棋或其他棋盘游戏。每个数据条目包括：puzzle_id（唯一标识符）、bitboard（三维int8列表，可能表示棋盘状态或位置信息）、mask（int8列表，可能用于标记有效区域或掩码）、label（int64，可能表示目标动作、分类或结果）、rating（int64，可能表示谜题难度或评分）、themes（字符串列表，可能表示谜题的主题或类别）。数据集分为训练集（4,663,652个示例）、测试集（582,957个示例）和验证集（582,956个示例），用于机器学习任务如分类或预测。

This dataset contains a series of puzzle or game-related data, potentially involving chess or other board games. Each data entry includes: puzzle_id (unique identifier), bitboard (a 3D int8 list, possibly representing board state or position information), mask (an int8 list, possibly used for marking valid areas or masks), label (int64, possibly indicating target action, classification, or outcome), rating (int64, possibly representing puzzle difficulty or rating), themes (a string list, possibly indicating puzzle themes or categories). The dataset is split into training set (4,663,652 examples), test set (582,957 examples), and validation set (582,956 examples), intended for machine learning tasks such as classification or prediction.