TIGER-Lab/ImagenWorld
收藏资源简介:
ImagenWorld是一个大规模的基准数据集,旨在评估在真实的多模态场景中的图像生成和编辑模型。它涵盖了六个不同的任务和六个内容领域,并以Parquet格式提供,包括结构化的元数据、文本提示和嵌入的参考图像。数据集还包括对‘训练’分割的人类注释。README还提供了使用🤗 datasets库加载数据集的信息,并为那些希望在研究中使用数据集的人提供了引用。
ImagenWorld is a large-scale benchmark designed to evaluate image generation and editing models in realistic multimodal scenarios. It spans six diverse tasks and six content domains, providing a unified framework for assessing model compositionality, instruction following, and multimodal capability. The dataset is provided in Parquet format, which includes structured metadata, text prompts, and embedded reference images for each task. It also includes human annotations for the train split. The README provides information on how to load the dataset using the 🤗 datasets library, and a citation for those who wish to use the dataset in their research.




