One-Gen Dataset

该数据集用于训练OneDiffusion模型，支持多种图像合成和理解任务，融合了多样化的内容。它不仅包含了如PixelProse、Unsplash、Coyo、JourneyDB等公共数据集，还包括了我们内部生成的多种数据集，如深度图、人体姿态和语义分割等。该数据集的规模宏大，既包括了公共数据集，也包含了一个拥有1000万个样本的大型内部合成数据集。它的任务涵盖了多种条件下的图像生成和理解。

This dataset is developed for training the OneDiffusion model, which supports a wide range of image synthesis and understanding tasks and integrates diversified content. It encompasses multiple public datasets including PixelProse, Unsplash, Coyo, and JourneyDB, as well as several internally generated datasets such as those for depth maps, human poses, and semantic segmentation. With a considerable scale, this dataset includes both public datasets and a large internal synthetic dataset containing 10 million samples. Its covered tasks span image generation and understanding under diverse conditions.