wulanbhai/subset

LaTeX公式数据集的桶索引，包含552,340个样本，按照80%训练集和20%测试集的比例随机分割（随机种子为42）。数据集根据公式长度和结构分为多个桶，如单行公式、多行公式、矩阵结构等，并保持了每个桶在训练集和测试集中的比例。Bucket indices for the LaTeX formulas dataset, containing 552,340 samples, split into 80% train set and 20% test set with a random seed of 42. The dataset is categorized into multiple buckets based on formula length and structure, such as single-line formulas, multi-line formulas, and matrix structures, maintaining the ratio of each bucket in both the train and test sets.