遇见数据集

bevaya/ENTRANT

收藏
Hugging Face2026-05-20 更新2026-07-22 收录
官方服务:

资源简介:

ENTRANT数据集是一个包含670万张结构化财务表格的大型数据集,这些表格提取自约33万份SEC EDGAR文件,涵盖10种文件类型(如年度报告、季度报告等)。该数据集用于表格问答和文本分类任务,专注于金融领域,基于SEC EDGAR公开文件构建,并遵循CC-BY-4.0许可证。数据集包括多个配置,每个配置对应一种SEC文件类型,并提供了表格ID、文件类型、标题、行列数、HTML渲染和单元格结构等详细元数据。

ENTRANT is a large dataset of 6.7 million structured financial tables extracted from approximately 330,000 SEC EDGAR filings spanning 10 filing types. It is designed for table-question-answering and text-classification tasks, focusing on the financial domain, and is built from publicly available SEC EDGAR documents under the CC-BY-4.0 license. The dataset includes multiple configurations, each corresponding to a specific SEC filing type, with detailed metadata such as table ID, filing type, title, row and column counts, HTML rendering, and cell structures.

提供机构:
bevaya
二维码
社区交流群
二维码
科研交流群
商业服务