1、配置依赖
$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
$ sudo apt-get install --no-install-recommends libboost-all-dev
2、ATLAS（ BLAS ）
$ sudo apt-get install libatlas-base-dev
3、Opencv
详细请见：https://blog.csdn.net/qq_33838170/article/details/83464525
4、Protobuf
$ git clone https://github.com/google/protobuf
$ sudo apt-get remove libprotobuf-dev
$ sudo apt-get install autoconf automake libtool curl make g++ unzip
$ cd protobuf
$ ./autogen.sh
$ ./configure --prefix=/usr
$ make
$ //make check
$ sudo make install
$ sudo ldconfig
卸载
which protoc
rm /usr/local/bin/protoc
然后删除文件夹
5、glog
$ wget https://github.com/google/glog/archive/v0.3.3.tar.gz
$ tar zxvf v0.3.3.tar.gz
$ cd glog-0.3.3
$ ./configure
$ make -j8
$ sudo make install -j8
6、gflags
$ wget https://github.com/schuhschuh/gflags/archive/master.zip
$ unzip master.zip
$ cd gflags-master
$ mkdir build && cd build
$ export CXXFLAGS="-fPIC" && cmake .. && make VERBOSE=1
$ make -j8
$ sudo make install -j8
7、 lmdb
$ git clone https://github.com/LMDB/lmdb
$ cd lmdb/libraries/liblmdb
$ make -j8
$ sudo make install -j8
8、Makefile.config配置
$ git clone https://github.com/BVLC/caffe
$ cd caffe
$ cp Makefile.config.example Makefile.config
$ sudo gedit Makefile.config
根据需要修改相关内容（去掉注释）
#CPU_ONLY := 1
#USE_CUDNN := 1
#OPENCV_VERSION := 3
#WITH_PYTHON_LAYER := 1
为避免发生hdf5相关错误，在Ubuntu16.04及以上版本需要将
#Whatever else you find you need goes here.
下面的
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib
替换为
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial
#CUSTOM_CXX := g+±4.8
9、Makefile配置
将
NVCCFLAGS +=-ccbin=$(CXX) -Xcompiler-fPIC $(COMMON_FLAGS)
替换为
NVCCFLAGS += -D_FORCE_INLINES -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)
10、编译
$ make all -j8//-j8是多核编译加快速度
$ make test -j8
$ make runtest -j8
11、愉快的debug
1、nvcc fatal : Unsupported gpu architecture ‘compute_20’
解决办法：
在 Makefile.config中去掉以下两行（由于是9以上版本、仔细看Makefile）
-gencode arch=compute_20,code=sm_20 \
-gencode arch=compute_20,code=sm_21 \
2、/usr/bin/ld: cannot find -lopencv_imgcodecs(或其他类似)
解决办法：
在Makefile最后加上
LIBRARIES += glog gflags protobuf leveldb snappy
lmdb boost_system boost_filesystem hdf5_hl hdf5 m
opencv_core opencv_highgui opencv_imgproc opencv_imgcodecs opencv_videoio
保存
make clean 再编译
3、collect2: error: ld returned 1 exit status
解决办法：
/usr/include/c++/4.8/bits/c++0x_warning.h:32:2: error: #error This file requires compiler and library support for the ISO C++ 2011 standard. This support is currently experimental, and must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
以上是原始的编译方式在Ubuntu18.04上的作死尝试，其实在Ubuntu18.04上可以用命令行安装预编译的CAFFE： sudo apt-get install caffe-cpu/cafe-cuda/caffedoc/caffe-tools-cpu/caffe-tools-cuda
caffeine 好像是一个java的高性能缓存库
只在caffedoc目录下找到了包括model、sample在内的相关文件，caffecpu内没有。
命令行运行caffe device_query test time train等命令出现错误：Check failed: FLAGS_model.size() > 0(0 vs. 0) Need a model definition to time. 暂未处理