ArkAiLab-Adl/Nexora-vision-dataset-v1-medium

Nexora Vision Dataset v1 Medium是一个包含5,231张高质量AI生成图像的数据集，专为生成式AI系统的研究、训练和评估而设计。数据集涵盖了电影、现实、动漫、科幻和艺术等多种风格，适用于图像生成、扩散模型训练、多模态学习等任务。数据集结构包括images、parquet等文件夹，图像格式为JPG，分辨率高。数据集布局和脚本使用MIT许可证，图像则遵循Pixabay许可证。

The Nexora Vision Dataset v1 Medium is an expanded release containing 5,231 high-quality AI-generated images built for serious research, training, and evaluation of generative AI systems. It offers a large-scale curated collection of cinematic, realistic, anime, sci-fi, and artistic images, ideal for training diffusion models, style extraction, computer vision experiments, and more. The dataset is structured with images and parquet folders, featuring high-resolution JPG images. The dataset layout and scripts are under the MIT License, while the images follow the Pixabay License.