guanweifan/efficient-vla-extracts

该数据集是`efficient-vla-wiki`项目的解析衍生品，用于支持论文查找、章节和图表定位、证据回溯、查询时报告生成以及维基维护。数据集包含从arXiv收集的论文，并通过`pdftotext`、`pdfinfo`、`Docling`和`Marker`等工具生成的解析文件。数据集结构包括`meta`和`parses`目录，其中`parses`目录下包含每篇论文的解析文件，如文本、边界框HTML、文档标记文件、JSON文件以及生成的图像和图表。使用和再分发需遵守每篇源论文的许可协议。This dataset is the parser-derived artifacts for the `efficient-vla-wiki` project, used for paper lookup, section and figure localization, evidence backtracking, query-time report generation, and wiki maintenance. It contains papers collected from arXiv, parsed using tools like `pdftotext`, `pdfinfo`, `Docling`, and `Marker`. The dataset structure includes `meta` and `parses` directories, with `parses` containing files such as text, bounding box HTML, document markup files, JSON files, and generated images and figures. Reuse and redistribution are subject to the license of each source paper.