Computer Vision | 最近的学习路线

Vision Transformer

paper：[2010.11929] An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale (arxiv.org)

bilibili：ViT论文逐段精读【论文精读】_哔哩哔哩_bilibili

code：WZMIAOMIAO/deep-learning-for-image-processing (github.com)

DeiT

paper：[2012.12877] Training data-efficient image transformers & distillation through attention (arxiv.org)

bilibili：DeiT：注意力Attention也能蒸馏 - 知乎 (zhihu.com)

code：同ViT

MAE

paper：[2111.06377] Masked Autoencoders Are Scalable Vision Learners (arxiv.org)

bilibili：MAE 论文逐段精读【论文精读】_哔哩哔哩_bilibili

bilibili：43、逐行讲解Masked AutoEncoder(MAE)的PyTorch代码_哔哩哔哩_bilibili

code：facebookresearch/mae

MoCo

paper：[1911.05722] Momentum Contrast for Unsupervised Visual Representation Learning (arxiv.org)

code：facebookresearch/moco

网上似乎关于MoCo的代码解读并不多，之后如果有时间可能录一份MoCo代码详解视频(如有)

Swim Transformer

......