Nayana-cognitivelab/nayana-beir-eval-multilang_v5

该数据集包含三个部分：文档集合(corpus)、查询与文档的相关性(qrels)、查询信息(queries)。文档集合中包含文档ID和图像，相关性部分包含查询ID、文档ID、是否可回答的标记、答案文本和分数，查询信息部分包含查询ID、查询文本、语言类型、GPT-4的推理过程和答案文本。数据集提供了测试集分割，适用于测试相关算法和模型。

The dataset consists of three parts: a corpus of documents, relevance information between queries and documents, and query details. The corpus includes document IDs and images, the relevance part includes query IDs, document IDs, answerability labels, answer text, and scores, while the query details part includes query IDs, query text, language type, GPT-4 reasoning process, and answer text. The dataset provides a test split for evaluating algorithms and models.