PictoViLT/merged_CG_L4_T

这是一个包含文本和图像数据的混合型数据集，适用于需要同时处理文本和图像的任务。数据集分为训练集和测试集，提供了文本的token信息以及图像的像素值信息。每个样本还包含了一些元数据信息，如块索引和是否仅包含被遮蔽的图像标记等。

This is a mixed-type dataset containing both text and image data, suitable for tasks that require simultaneous processing of text and images. The dataset is divided into training and test sets, providing token information for text and pixel values for images. Each sample also includes metadata such as chunk index and whether it only includes masked image tokens.