MBZUAI/ViMUL-Bench
收藏Hugging Face2025-06-11 更新2025-07-05 收录
下载链接：
https://hf-mirror.com/datasets/MBZUAI/ViMUL-Bench
下载链接
链接失效反馈官方服务：
资源简介：
ViMUL-Bench是一个包含14种语言和15个类别的多语言多模态视频数据集，用于评估多语言视频语言模型。数据集包含生活方式、节日、食物、仪式、当地地标、文化名人等文化多样性类别。视频时长分为短、中、长三种，所有样本均经过母语人士手动验证，确保质量。
ViMUL-Bench is a multilingual multimodal video dataset containing 14 languages and 15 categories, designed for evaluating multilingual video language models. The dataset includes diverse cultural categories such as lifestyles, festivals, foods, rituals, local landmarks, and cultural personalities. Video durations are categorized as short, medium, and long, with all samples manually verified by native speakers for quality assurance.
提供机构：
MBZUAI
