vidore/tatdqa_test_beir

这是一个包含文本和图像数据的多配置数据集，分为corpus、qrels和queries三个部分。corpus部分包含文本id和图像，qrels部分包含查询id、文本id和相关性分数，queries部分包含查询id和查询文本。数据集适用于文本图像检索任务，并提供测试分割的数据。

This is a multi-config dataset containing text and image data, divided into three parts: corpus, qrels, and queries. The corpus part includes text ids and images, the qrels part includes query ids, text ids, and relevance scores, and the queries part includes query ids and query texts. The dataset is suitable for text-image retrieval tasks and provides data for the test split.