nanonets/long_sparse_structured_table

这是一个合成的数据集，用于创建具有以下特征的表格：空白单元格比例在40%到70%之间（稀疏），行与列之间有清晰的分隔（结构化），表格行数在15到30之间，列数在7到15之间（较长）。This dataset is generated synthetically to create tables with the following characteristics: an empty cell percentage within the range [40,70] (Sparse), a clear separator between rows and columns (Structured), and a number of rows between 15 and 30, and a number of columns between 7 and 15 (Long).