mixedbread-ai/vidore-infovqa_test_subsampled

该数据集由三个子数据集组成，分别是corpus、default和queries。corpus子数据集包含图片和相关ID，default子数据集包含查询ID、语料库ID和分数，而queries子数据集包含查询ID和查询文本。每个子数据集都提供了训练集分割。The dataset consists of three sub-datasets: corpus, default, and queries. The corpus sub-dataset includes images and associated IDs, the default sub-dataset contains query IDs, corpus IDs, and scores, while the queries sub-dataset includes query IDs and query texts. Each sub-dataset provides a training set split.