Auto-captions on GIF
收藏arXiv2020-07-06 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2007.02375v1
下载链接
链接失效反馈官方服务：
资源简介：
一个用于通用视频理解预训练的新的大型数据集，通过自动提取和筛选来自数十亿网页的视频标题注释创建所有视频-句子对。该数据集可用于预训练通用特征表示或编码器-解码器结构，用于视频字幕，以及其他下游任务（如视频中的句子定位、视频问答等）。
A novel large-scale dataset for general video understanding pre-training. All video-sentence pairs in this dataset are generated through automatic extraction and filtering of video title annotations from billions of web pages. This dataset can be used to pre-train general feature representations or encoder-decoder architectures for video captioning, as well as other downstream tasks such as sentence localization in videos and video question answering.
创建时间：
2020-07-06
