Nano1337/FPSv2-scores-natural-VQA-not-FPSv2

该数据集包含100个训练样本，每个样本包括一个唯一的用户ID（uid）、一张图片（image）、一段描述（caption）、FPSv2评分、VQA评分、FPSv2百分位数、VQA百分位数以及百分位数差异。这些特征可能用于图像和文本的联合分析，如图像标注或视觉问答系统的训练和评估。This dataset contains 100 training samples, each including a unique user ID (uid), an image, a caption, FPSv2 score, VQA score, FPSv2 percentile, VQA percentile, and percentile difference. These features may be used for joint analysis of images and text, such as image captioning or training and evaluation of visual question answering systems.