Nayana-cognitivelab/nayana-beir-eval-multilang_v9

该数据集包含三个部分：文档集合(corpus)、查询与相关文档(qrels)、查询信息(queries)。文档集合包含文档的标识和图像信息，查询与相关文档部分包含查询与文档的关系信息，包括是否为答案以及答案内容。查询信息部分包含查询的文本内容、语言、GPT-4推理结果和答案。所有数据均针对测试集进行分割。

The dataset consists of three parts: document collection (corpus), query-relevant document (qrels), and query information (queries). The document collection includes document identifiers and image information. The query-relevant document part contains information about the relationship between queries and documents, including whether it is an answer and the answer content. The query information part includes the text content of the query, language, GPT-4 reasoning results, and answers. All data are split for the test set.