Intelligent-Internet/pd12m

PD12M数据集是一个经过筛选的Public Domain 12M图像集合，包含EXIF数据和通过SigLIP 2进行预处理和特征提取得到的向量嵌入。数据集结构包括图像ID、URL、字幕、原始和处理后图像的尺寸、宽高比、EXIF数据、元数据、创建和更新时间、来源、向量嵌入、原始来源和许可证信息。

The PD12M dataset is a curated collection of Public Domain 12M images, including EXIF data and vector embeddings obtained through preprocessing and feature extraction with SigLIP 2. The dataset structure includes image ID, URL, captions, original and processed image dimensions, aspect ratio, EXIF data, metadata, creation and update times, source, vector embedding, original source, and license information.