bliberatori/mapillary_pair_generated

该数据集包含多个图像数据划分，主要特征包括图像、城市、图像ID、变体和元数据字段。数据划分包括训练集及其变体，如train、train_overlap、train_outpaint_homography等，每个划分都有特定的示例数量和字节大小。数据集总大小约为8.93 GB，下载大小约为10.40 GB。可能用于图像处理、城市视觉分析或生成模型训练等任务，但具体用途和背景未在README中明确说明。This dataset includes multiple splits of image data, with key features such as image, city, image_id, variation, and metadata. The splits consist of training sets and variants like train, train_overlap, train_outpaint_homography, etc., each with specified numbers of examples and byte sizes. The total dataset size is approximately 8.93 GB, with a download size of about 10.40 GB. It may be intended for tasks such as image processing, urban visual analysis, or generative model training, but the specific purpose and background are not explicitly detailed in the README.