jun-2018/sampling_multimodal_doc_pages_v4

该数据集是一个包含文档信息的集合，其中包括文档ID、分割方式、文档类型等元数据，以及文档的原始PDF路径、标签路径、页面图像路径等文件路径信息。每个文档还有页码、页面OCR文本、页面PDF文本等内容。此外，数据集还详细记录了视觉元素和表格元素的信息，如OCR内容、路径、PDF内容、类型、Markdown格式、令牌数量等。数据集分为验证集和训练集，分别用于模型验证和训练。

This dataset is a collection of document information, including metadata such as document ID, split type, document type, and file path information such as original PDF path, label path, and page image path. Each document also contains page number, page OCR text, page PDF text, etc. In addition, the dataset records detailed information of visual elements and table elements, such as OCR content, path, PDF content, type, Markdown format, token count, etc. The dataset is divided into validation set and training set for model validation and training respectively.