boombung的博客

《Self-supervised Video Representation Learning Using Inter-intra Contrastive Framework》论文阅读笔记论文地址：https://arxiv.org/pdf/2008.02531.pdf代码地址：https://github.com/BestJuly/IIC目录引言一、文章解析1.1 backbone1.2 inputs1.3 contrastive learning1.4 joint representation1.5

《Listen to look：Action recognition by previewing audio》论文阅读笔记引言视频冗余IMGAUD2VID 师生蒸馏框架current approachesauthorIMGAUD-SKIMMING attention-based LSTM网络LSTMQueryscore预测总结论文地址：https://arxiv.org/abs/1912.04487代码地址：https://github.com/facebookresearch/Listen-to-Lo