PictoViLT/CG_L3_img_F

该数据集包含用于训练的文本和图像数据，特征字段涵盖了文本处理所需的输入ID、注意力掩码、令牌类型ID和标签，以及图像处理所需的像素值和遮蔽索引。元数据提供了关于数据集的额外信息，如是否仅包含遮蔽token的图片、数据集的逻辑名称和遮蔽图片的方式。训练集共有617个样本，总大小为1,095,181,943字节。

This dataset includes text and image data for training, with feature fields covering input IDs, attention masks, token type IDs, and labels required for text processing, as well as pixel values and masked indices for image processing. The metadata provides additional information about the dataset, such as whether it includes only masked tokens images, the logical name of the dataset, and the method of masking images. The training set consists of 617 samples with a total size of 1,095,181,943 bytes.