TencentARC/TimeLens-100K

TimeLens-100K是一个大规模、多样化和高质量的视频时间定位训练数据集，用于训练TimeLens模型。数据集包含约20K视频和约100K标注，平均每个视频有约5个标注。视频来源包括DiDeMo、QuerYD、HiREST、CosMo-Cap和InternVid-VTime等多个数据集。TimeLens-100K is a large-scale, diverse, and high-quality training dataset for video temporal grounding. It was proposed in our paper [TimeLens: Rethinking Video Temporal Grounding with Multimodal LLMs](TODO) and used for training [TimeLens models](https://github.com/TencentARC/TimeLens#-timelens-models). The dataset contains ~20K videos and ~100K annotations, with an average of ~5 annotations per video. Video sources include DiDeMo, QuerYD, HiREST, CosMo-Cap, and InternVid-VTime.