qz1188/snli-ve

该数据集是一个多模态数据集，包含图像和文本对，基于Flickr30K构建。每个样本包括Flickr30K_ID（标识符）、gold_label（黄金标签，可能用于分类或评估）、image（图像数据）和sentence（与图像相关的描述性句子）。数据集分为训练集、验证集和测试集，分别用于模型训练、验证和测试，总样本量约为565,286个。This dataset is a multimodal dataset comprising image-text pairs, which is constructed based on the Flickr30K dataset. Each sample contains Flickr30K_ID (identifier), gold_label (a gold standard label that can be used for classification or evaluation), image (image data), and sentence (a descriptive sentence related to the corresponding image). The dataset is split into training, validation, and test sets, which are used for model training, validation, and testing respectively, with a total of approximately 565,286 samples.