lightonai/veracier-industries

EDiTh（企业数字孪生基准）是一个开放的基准数据集，用于评估企业搜索和RAG系统。它包含1004个PDF文档（约1.7GB），涵盖合同、报告、政策等多种类型，支持法语、英语、德语、意大利语和西班牙语等多种语言。数据集模拟了一个虚构的法国工业集团Véracier Industries S.A.及其子公司的文档，包括搜索able、扫描和混合格式的PDF。数据集还包含36个评估用例，用于测试检索和问答系统的性能。所有文档均为合成，不包含真实信息。

EDiTh (Enterprise Digital Twin Benchmark) is an open benchmark for evaluating enterprise search and RAG systems. It includes 1,004 PDF documents (~1.7 GB) such as contracts, reports, and policies, in multiple languages (French, English, German, Italian, Spanish). The dataset simulates documents from a fictional French industrial group, Véracier Industries S.A., and its subsidiaries, with PDFs in searchable, scanned, and mixed formats. It also features 36 evaluation use cases to test retrieval and question-answering performance. All documents are synthetic and do not contain real information.