izhx/UMRB-EncyclopediaVQA

该数据集包含三个配置：query、corpus和qrels。qrels配置用于存储查询与语料库之间的相关性评分，包含查询ID、Q0、语料库ID和分数特征，测试集有4887个样本。corpus配置用于存储语料库数据，包含ID、模态、文本和图像特征，语料库集有68313个样本。query配置用于存储查询数据，包含ID、模态、文本和图像特征，测试集有3743个样本。数据集主要用于信息检索和多模态检索任务。

The dataset contains three configurations: query, corpus, and qrels. The qrels configuration is used to store relevance scores between queries and the corpus, containing features such as query-id, Q0, corpus-id, and score, with 4887 examples in the test split. The corpus configuration is used to store corpus data, containing features such as id, modality, text, and image, with 68313 examples in the corpus split. The query configuration is used to store query data, containing features such as id, modality, text, and image, with 3743 examples in the test split. The dataset is primarily used for information retrieval and multimodal retrieval tasks.