FoteiniTag/modified_finepdfs_ell_grek_pages

这是一个多模态数据集，包含图像和对话文本数据。每个样本由以下字段组成：图像（image字段）、对话列表（conversations字段，其中每个对话包含发言者from和内容value）、PDF标识符（pdf_id）、页码（page_num）、总页数（num_pages）、来源URL（url）和源行索引（source_row_idx）。数据集仅包含训练拆分（train），共有53个样本，总大小约22.6MB。该数据集可能用于文档理解、视觉问答或基于PDF页面图像的对话生成任务。

This is a multimodal dataset containing image and conversational text data. Each sample consists of the following fields: image, conversations (a list where each conversation includes from and value), pdf_id, page_num, num_pages, url, and source_row_idx. The dataset only includes a train split with 53 examples and a total size of approximately 22.6MB. It is likely designed for document understanding, visual question answering, or dialogue generation tasks based on PDF page images.