Condensed Movies

该数据集旨在验证ContextIQ在视频检索任务中的性能，并与TwelveLabs以及其他模型进行了比较。此外，该数据集的质量控制采用了基于投票的系统，由标注员进行人工验证。所涉及的任务是文本到视频的检索。

This dataset is designed to validate the performance of ContextIQ in video retrieval tasks, and compare it with TwelveLabs and other models. Furthermore, a voting-based quality control system is adopted for this dataset, where manual verification is carried out by annotators. The task involved in this dataset is text-to-video retrieval.