jhkwak-bp/lvis-label

这是一个包含图片及其注释信息的图像数据集。数据集中的每条记录都包括一个注释id、图片id、分类id、图片详细信息（如宽度、高度、捕获日期、flickr链接、COCO链接、许可证信息、负分类id、非详尽分类id）、标注信息（如边界框、区域、分割）、分类详细信息（如名称、定义、同义词、同义词集、频率、图片数量、实例数量）以及数据集划分信息（训练集和验证集）。图片数据集总共包含1141万多个示例，分为训练集和验证集两部分，分别有1270141和244707个示例。

This is an image dataset containing images and their annotations. Each record in the dataset includes an annotation id, image id, category id, image details (such as width, height, capture date, flickr URL, COCO URL, license information, negative category ids, non-exhaustive category ids), annotation details (such as bounding box, area, segmentation), category information (such as name, definition, synonyms, synset, frequency, image count, instance count), and dataset split information (train and validation sets). The image dataset contains over 11.41 million examples, divided into two parts: training set and validation set, with 1,270,141 and 244,707 examples respectively.