OVEN

该数据集包含图像三元组、查询文本和文档段落，共包含3,500篇文档和37,600张相关图片。平均每篇文档有10.7对配对图像。规模上，数据集涵盖了3,500篇文档和37,600张图片。该数据集的任务是图像-文档检索。

This dataset comprises image triplets, query texts, and document paragraphs, containing a total of 3,500 documents and 37,600 relevant images. On average, each document has 10.7 pairs of matched images. In terms of scale, the dataset covers 3,500 documents and 37,600 images. The task of this dataset is image-document retrieval.