nanonets/small_sparse_unstructured_table

这个数据集是合成的，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），行与列之间没有分隔符（非结构化），表格大小为4到10行，2到6列。数据集适用于表格问答任务，并包含光学字符识别、表格处理和智能文档处理的相关数据。

This dataset is synthetically generated to create tables with the following characteristics: an empty cell percentage within the range of [40,70] (Sparse), no separator between rows and columns (un-structured), and table size ranging from 4 to 10 rows and 2 to 6 columns. The dataset is designed for table question answering tasks and includes data related to optical character recognition, table processing, and intelligent document processing.