naist-nlp/BQA
收藏Hugging Face2025-06-05 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/naist-nlp/BQA
下载链接
链接失效反馈官方服务：
资源简介：
BQA数据集包含7,632个5-10秒的短视频（25帧/秒），每个视频都有关于人体语言的元数据（性别、年龄、种族）以及26个情感标签。数据集的创建包括使用Gemini-1.5-pro提取答案选项、生成问题、评估潜在的危害以及分配难度标签四个步骤。该数据集用于评估视频大型语言模型的性能，并揭示了模型在不同人口统计信息上的理解差异。
The BQA dataset consists of 7,632 short videos (5-10 seconds, 25 fps) depicting human body language, each with metadata (gender, age, ethnicity) and 26 emotion labels per video. The creation of the BQA involves four steps using Gemini-1.5-pro: extracting answer choices, generating questions, evaluating potential harm, and assigning difficulty labels. The dataset is used to evaluate the performance of video large language models and reveals disparities in understanding across demographics.
提供机构：
naist-nlp
