本文是对ImageNet 2015的冠军ResNet(Deep Residual Networks)以及目前围绕ResNet这个工作研究者后续所发论文的总结，主要涉及到下面5篇论文。     阅读全文

1. Deep Residual Network论文下载 在2015年的ImageNet比赛中，MSRA的He Kaiming用deep residual network取得了惊人的成果，不仅在多个task上排名榜首，而且远超第二名好多好多。这极大的引起了我对Deep Residual Network的好奇心，想要玩一玩Deep Residual Network。     阅读全文

转载请注明！！！ Sometimes we want to implement new layers in Caffe for specific model. While for me, I need to Implement a L2 Normalization Layer. The benefit of applying L2 Normalization to the data is obvious. The author of Caffe has already wrote methods to add new layers in Caffe in the Wiki. This is the Link     阅读全文