nanonets/nn-auto-bench-ds

nn-auto-bench-ds是一个为关键信息提取（KIE）设计的基准数据集。该数据集包含1000个文档，分为发票、收据、护照和银行对账单等类型。文档主要是英文的，但也有德语和阿拉伯语的文档。每个文档都标注了用于关键信息提取和特定任务的信息。该数据集可用于计算大型语言模型在KIE任务上的oneshot性能。

nn-auto-bench-ds is a benchmark dataset designed for key information extraction (KIE). The dataset includes 1,000 documents, categorized into types such as invoices, receipts, passports, and bank statements. The documents are primarily in English, with some also in German and Arabic. Each document is annotated with information for key information extraction and specific tasks. The dataset can be used to compute the oneshot performance of large language models on KIE tasks.

搜集汇总

背景与挑战

背景概述

该数据集是用于关键信息提取（KIE）的基准数据集，包含1,000个文档，涵盖发票、收据、护照和银行对账单四种类型，主要以英文为主，部分支持德文和阿拉伯文。它整合了多个开源数据源，旨在评估大型语言模型在文档理解任务中的一次性性能，适用于自然语言处理研究。