MMNER

该数据集是一个大规模的多语言和多模态命名实体识别（NER）数据集，包含四种语言（英语、法语、德语和西班牙语）的图像-文本对。该数据集涵盖四个类别（人物、地点、组织和杂项），共89,019个实体，并且已经通过高可靠性的手动标注（Cohen's kappa系数为0.96）。其规模达到33,965个图像-文本对，适用于多语言和多模态命名实体识别（Mmner）任务。

This dataset is a large-scale multilingual and multimodal named entity recognition (NER) dataset containing image-text pairs across four languages: English, French, German and Spanish. It covers four categories (person, location, organization and miscellaneous) with a total of 89,019 entities, and has been manually annotated with high reliability (Cohen's kappa coefficient of 0.96). Comprising 33,965 image-text pairs, this dataset is suitable for multilingual and multimodal named entity recognition (Mmner) tasks.