vidore/docvqa_test_subsampled_beir

这是一个包含图片和对应查询及其相关性的数据集，分为corpus、qrels和queries三个部分。corpus部分包含图片数据，qrels部分包含查询ID、文档ID和分数，用于评估相关性，queries部分包含查询内容。数据集包含一个测试集，共有500个示例。

This dataset consists of images and corresponding queries with their relevance, divided into three parts: corpus, qrels, and queries. The corpus part contains image data, the qrels part contains query IDs, document IDs, and scores for relevance evaluation, and the queries part contains the query content. The dataset includes a test split with a total of 500 examples.