G4KMU/t2-ragbench
收藏Hugging Face2025-06-17 更新2025-11-01 收录
下载链接：
https://hf-mirror.com/datasets/G4KMU/t2-ragbench
下载链接
链接失效反馈官方服务：
资源简介：
T2-RAGBench是一个金融领域的文本和表格基准数据集，用于评估检索增强生成技术在处理金融文档时的表现。它由32,908个独立于上下文的问答对和超过9000个文档组成，这些文档来源于四个精细数据集：FinQA、ConvFinQA、VQAonBD和TAT-DQA。数据集支持问答、基于表格的问答和检索增强生成任务，并包含了原始的PDF文件以供直接使用。
T2-RAGBench is a text-and-table benchmark dataset in the financial domain designed for evaluating the performance of Retrieval-Augmented Generation (RAG) techniques on financial documents. It consists of 32,908 context-independent question-answer pairs and over 9,000 documents from four refined datasets: FinQA, ConvFinQA, VQAonBD, and TAT-DQA. The dataset supports Question Answering, Table-based Question Answering, and Retrieval-Augmented Generation tasks, and includes original PDF files for direct use.
提供机构：
G4KMU
