bigdata-pw/Diffusion1B

Diffusion1B数据集包含约12亿张由扩散模型生成的图像，每张图像的生成细节如提示词、使用的模型、匿名用户信息、创建日期和图像URL等都被记录。该数据集是bigdata-pw/leonardo和bigdata-pw/playground两个数据集的组合。数据集由hlky策划，采用Open Data Commons Attribution License (ODC-By) v1.0授权。The Diffusion1B dataset contains approximately 1.2 billion images generated by diffusion models, with details including the prompts and models used, anonymized user information, creation date, and URL to the image. Curated by hlky, the dataset is released under the Open Data Commons Attribution License (ODC-By) v1.0. The dataset structure details multiple attributes for each image, such as unique identifier, user identifier, creation time, base model, custom model, prompt, negative prompt, sampler, seed value, number of steps, classifier-free guidance scale, width and height of the image, URL to the image, and like count.