mteb/JinaVDRAutomobileCatelogRetrieval

这是一个基于LLM生成的查询来检索汽车营销文档的日语数据集，包含图像和文本特征，分为corpus、qrels和queries三个配置。corpus配置包含图像和ID，qrels配置包含查询ID、语料库ID和分数，queries配置包含ID、文本和模态。数据集用于视觉文档检索、图像到文本和文本到图像的任务。

This is a Japanese dataset for retrieving automobile marketing documents based on LLM-generated queries, containing image and text features, divided into three configurations: corpus, qrels, and queries. The corpus configuration includes images and IDs, the qrels configuration includes query IDs, corpus IDs, and scores, and the queries configuration includes IDs, text, and modality. The dataset is used for visual document retrieval, image-to-text, and text-to-image tasks.