05 2021 档案

摘要：Data Uncertainty Learning in Face Recognition 2021.5.25 CVPR20 Abstract 对于有噪声的图片，对于数据的不确定性非常重要，但很少有研究面部识别。之前的PFE将每个面部的嵌入作为高斯分布来考虑不确定性。它使用先前已有模型的固定特征作为 阅读全文

摘要：Self-distillation with Batch Knowledge Ensembling Improves ImageNet Classification 2021.5.13 Project Page: https://geyixiao.com/projects/bake https:// 阅读全文

摘要：Distill on the Go: Online knowledge distillation in self-supervised learning 2021.5.13 Learning from Limited or Imperfect Data (L2ID) Workshop @ CVPR 阅读全文

摘要：Beyond Self-attention: External Attention using Two Linear Layers for Visual Tasks 2021.5.12 Under review https://arxiv.org/abs/2105.02358 Introductio 阅读全文

摘要：Decoupled Dynamic Filter Networks 2021.5.6 CVPR 2021 https://arxiv.org/abs/2104.14107 Introduction 卷积缺点在于：内容不变，计算量高 动态filter可以根据内容自适应，但是会提高计算量。depth-w 阅读全文

摘要：Accelerate CNNs from Three Dimensions: A Comprehensive Pruning Framework 2021.5.10 ICML 2021 https://arxiv.org/abs/2010.04879 Introduction 大部分的剪枝pruni 阅读全文