Thursday, November 14, 2019

WRN

WRN

2019/11/12

前言:

WRN 是 Wide ResNet 的縮寫,論文名稱則是《Wide residual networks》。在 PreVGGNet 的作者嘗試加寬網路效果不大之後,增加了恆等映射的殘差網路使的加寬網路又成為可行。WRN 的作者發現,矮胖的殘差網路可以達到跟高瘦的殘差網路接近的效果。

-----



// Wide ResNet _ PyTorch

-----


// Review Wide Resnet

-----


// Multi-class Image Classification using deep convolutional networks on…

-----



Dropout in Original ResNet (Left) and Dropout in WRNs (Right)

// Review  WRNs — Wide Residual Networks (Image Classification)

-----


# WRN

-----


# WRN

-----


# WRN

-----


# WRN

-----


# WRN

-----

References

◎ 論文

# WRN
Zagoruyko, Sergey, and Nikos Komodakis. "Wide residual networks." arXiv preprint arXiv:1605.07146 (2016).
https://arxiv.org/pdf/1605.07146.pdf 

-----

◎ 英文參考資料  

Wide ResNet _ PyTorch
https://pytorch.org/hub/pytorch_vision_wide_resnet/

Review  WRNs — Wide Residual Networks (Image Classification)
https://towardsdatascience.com/review-wrns-wide-residual-networks-image-classification-d3feb3fb2004

Multi-class Image Classification using deep convolutional networks on…
https://www.slideshare.net/ashisface/multiclass-image-classification-using-deep-convolutional-networks-on-extremely-large-dataset

-----

◎ 韓文參考資料   

Wide Residual Networks
https://norman3.github.io/papers/docs/wide_resnet.html

Review Wide Resnet
https://www.slideshare.net/WoojinJeong5/wide-resnet

-----

◎ 簡體中文參考資料

WRNS:Wide Residual Networks 论文笔记 - XlyPb - CSDN博客
https://blog.csdn.net/wspba/article/details/72229177

[DL-架构-ResNet系] 004 WRN - 知乎
https://zhuanlan.zhihu.com/p/29681673

No comments: