精简CNN模型

精简CNN模型

一、MobileNet

MobileNet V1理论：Depthwise Convolution + Pointwise Convolution
https://blog.youkuaiyun.com/u014380165/article/details/72938047
MobileNet V2理论：Linear Bottlenecks + Inverted residuals
https://blog.youkuaiyun.com/u014380165/article/details/79200958

代码：https://www.cnblogs.com/ansang/p/9159440.html

二、SqueezeNet

https://www.jianshu.com/p/9a70942af217

三、SkipNet

https://www.jianshu.com/p/72bbd7d1d2de

四、ShuffleNet

ShuffleNet V1理论：https://www.jianshu.com/p/44db6d72d6eb
ShuffleNet V2理论：https://www.jianshu.com/p/71e32918ea0a

五、Xception

理论：Depthwise Convolution + Pointwise Convolution
https://blog.youkuaiyun.com/u014380165/article/details/75142710