abdulahmad0001/From

LLaVA-Video-178K数据集由Yuanhan Zhang、Jinming Wu和Wei Li策划，主要用于LLaVA-Video模型的训练。数据集包含178,510条字幕条目、960,792个开放式问答（QA）项目和196,198个多项选择QA项目。数据来源于五个主要来源：LLaVA-Video-178K、NeXT-QA、ActivityNetQA、PerceptionTest和LLaVA-Hound。数据集提供多种分割和配置，包括字幕、开放式问答和多项选择问答。数据集仅允许用于学术研究和教育目的。The LLaVA-Video-178K dataset is curated by Yuanhan Zhang, Jinming Wu, and Wei Li, primarily for training the LLaVA-Video model. It includes 178,510 caption entries, 960,792 open-ended QA (question and answer) items, and 196,198 multiple-choice QA items. The data is sourced from five primary sources: LLaVA-Video-178K, NeXT-QA, ActivityNetQA, PerceptionTest, and LLaVA-Hound. The dataset provides various splits and configurations, including captions, open-ended QA, and multiple-choice QA. The dataset is only allowed for academic research and education purposes.