helena-balabin/vg-captions-graphs

该数据集包含图片特征、图像图、动作图像图、空间图像图和原始文本句子。具体包含宽度、高度、URL、图像ID、COCO ID、Flickr ID等字段，以及图像图的边索引、节点数和补丁索引等结构信息。The dataset includes image features, image graphs, action image graphs, spatial image graphs, and raw text sentences. It contains fields such as width, height, URL, image ID, COCO ID, Flickr ID, and structural information such as edge indices, number of nodes, and patch indices for the image graphs.