64位win7下，vs2015编译、配置caffe

在vs2015配置caffe，摸索了很久才将caffe编译成功，安装、编译、环境配置过程整理：


1.下载并安装Anaconda2-5.0.1-Windows-x86_64.exe（Python2.7）或者安装python3.5。注意，暂时不支持Python3.6

下载地址：http://download.csdn.net/download/hhhhggggffff/10010450

2.安装CMake3.4.exe，配置系统环境（./CMake/bin）

3.安装cuda_8.0.61_windows.exe
下载地址：https://developer.nvidia.com/cuda-80-ga2-download-archive

4.下载cudnn-8.0-windows7-x64-v5.1.zip， 解压该文件，将文件夹bin,include,lib中的文件拷贝到C:\Program
Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0中对应的bin,include,lib文件夹中。
下载地址：http://download.csdn.net/download/demoscene/9754384

5.下载caffe-windows.zip，解压文件。在.\caffe-windows\scripts中，文本打开build_win.cmd，修改配置


（1）WITH_NINJA=0      使用cmake编译
     MSVC_VERSION=14    VS版本
     PYTHON_VERSION=2   python的版本



（2）删除以下内容



（3）依照（1）修改相应配置



6.编译
打开cmd命令行，将路径cd到.\caffe-windows下，将.\caffe-windows\scripts\
build_win.cmd拖拽到cmd中，并回车。自动下载一些文件或者库，并生成bulid文件夹。大概需要15分钟左右。

7.将libraries_v140_x64_py27_1.1.0.tar.bz2解压到build文件夹。并将./build/libraries/bin和./build/libraries/lib加到环境变量。

下载地址：https://github.com/willyd/caffe-builder/releases/download/v1.1.0/libraries_v140_x64_py27_1.1.0.tar.bz2


8.用vs2015打开.\caffe-windows\bulid\Caffe.sln，编译Debug/Release。可能是不需要再重新编译Release，cmake已经生成了。
9.环境变量：C:\Program Files\NVIDIA GPU Computing
Toolkit\CUDA\v8.0
10.vs2015配置caffe
包含目录：

D:\program\Caffe\caffe-windows\src

D:\program\Caffe\caffe-windows\build\libraries\include\boost-1_61

D:\program\Caffe\caffe-windows\build\libraries\include

D:\program\Caffe\caffe-windows\build\libraries\include\opencv

D:\program\Caffe\caffe-windows\include

D:\program\Caffe\caffe-windows\build

D:\program\Caffe\caffe-windows\build\include

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\include

库目录：

D:\program\Caffe\caffe-windows\build\lib\Release

D:\program\Caffe\caffe-windows\build\libraries\x64\vc14\lib

D:\program\Caffe\caffe-windows\build\libraries\x64\vc14\bin

D:\program\Caffe\caffe-windows\build\libraries\bin

D:\program\Caffe\caffe-windows\build\libraries\lib

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\lib\x64

D:\program\Anaconda2\libs

链接器-》输入：

caffe.lib
caffeproto.lib
boost_system-vc140-mt-1_61.lib
boost_thread-vc140-mt-1_61.lib
boost_chrono-vc140-mt-1_61.lib
boost_date_time-vc140-mt-1_61.lib
boost_atomic-vc140-mt-1_61.lib
boost_python-vc140-mt-1_61.lib
boost_filesystem-vc140-mt-1_61.lib
glog.lib
gflags.lib
shlwapi.lib
libprotobuf.lib
caffehdf5_hl.lib
caffehdf5.lib
lmdb.lib
ntdll.lib
leveldb.lib
snappy_static.lib
caffezlib.lib
opencv_highgui310.lib
opencv_imgcodecs310.lib
opencv_imgproc310.lib
opencv_core310.lib
libopenblas.dll.a
libboost_system-vc140-mt-1_61.lib
cudnn.lib
cublas.lib
cuda.lib
cublas_device.lib
cudadevrt.lib
cudart.lib
cudart_static.lib
cufft.lib
cufftw.lib
curand.lib
cusolver.lib
cusparse.lib
nppc.lib
python27.lib