ReadMe

./doc/ZTE_challenge_2019.pdf is the documentation of the methods I used in ZTE challenge preliminary

Firstly, you should add all of *.cu and *.cpp to src/caffe/layers, and add all of *.hpp to include/caffe/layers . Then， you need change the caffe.proto :

message ConvolutionParameter {
...
optional bool combine_relu = 20 [default = false]; // add this sentence
...
}

then，recompile Caffe。

How to use

If you want to combine ReLU layer with convolutional layer (only on .caffemodel TEST stage), just delete ReLU layer directly and set "combine_relu : true" [default=False]。

Some blogs about how to compile Caffe

On Ubuntu

Ubuntu 16.04 下用 cmake 安装 caffe

On windows

Windows 下用 build_win.cmd 直接编译CPU版caffe Windows 下用 build_win.cmd 直接编译 GPU 版caffe