2020 CVPR - Momentum Contrast for Unsupervised Visual Representation Learning. Kaiming He et al. 

2021 - VideoLT: Large-scale Long-tailed Video Recognition

论文中如何写伪代码