【CS231n winter2016 Lecture 6 (Training Neural Networks Part 3,update schemes/dropout and so on..)】
2016-11-29 18:09
295 查看
lecture 6 Neural Networks Part 3 Intro to ConvNets hd
outline:SGD performes not so good in practice.
Momentum update ,物理解释很不错
Nesterov Momentum update
and so on…
二次优化的方法,Hessian矩阵要求逆,尤其是高维的时候,所以impratical
改进的
ensemble
dropout(regularization)
see the slides for detail..
相关文章推荐
- 【CS231n winter2016 Lecture 5 (Training Neural Networks Part 1)】
- 【CS231n】Lecture 6:Training Neural Networks,Part 2
- char * a, char ** a, char * a[], char a[][], char * a[][], char ** a[][], char * a [][][], and so on
- 论文阅读《2018.2-Do deep nets really need weight decay and dropout》
- createDataSource, and add new rows on the datagrid to be filled out.
- Outlook 打开, 关闭变慢(Outlook 2007 so slow and not just me..check this out!)
- STL运用集合!(vector,map,list and so on)!!
- [Clojure] Data Collection and Data Analysis on the music of www.xiami.com - Part 5
- C# Disable CTRL-ALT-DEL, ALT-TAB, ALT-F4, Start Menu and so on…
- implement drag and drop on android
- 解决AndroidStudio添加ProjectLibary后在编译时遇到的各种问题之解决方式索引(finished with non-zero exit value and so on...)
- mdss 8x26/8x74 display debug and reading notes(vsync, interface and so on)
- char * a, char ** a, char * a[], char a[][], char * a[][], char ** a[][], char * a [][][], and so on
- 解决AndroidStudio添加ProjectLibary后在编译时遇到的各种问题之解决方式索引(finished with non-zero exit value and so on...)
- No update on April 11 and 12, 2015!
- Handbook on Information Technologies for Education and Training
- SVD, PCA and so on
- [翻译]So you want to learn WSE 3.0? A short primer on how and where to start
- no ip proxy-arp and so on
- char * a, char ** a, char * a[], char a[][], char * a[][], char ** a[][], char * a [][][], and so on