nanonets/long_dense_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：空单元格百分比在0到30之间（密集），行与列之间有清晰的分隔符（结构化），行数在15到30之间，列数在7到15之间（长表格）。该数据集适用于表格问答任务，并包含光学字符识别和智能文档处理相关的表格。

This dataset is generated synthetically to create tables with the following characteristics: an empty cell percentage within the range [0,30] (Dense), 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). The dataset is designed for table question answering tasks and includes tables related to OCR (Optical Character Recognition) and IDP (Intelligent Document Processing).