和TensorFlow对应的是Theano，Torch；
Caffe专精于图像处理，Caffe方便，更快入门上手；
在通用的DL task上，Caffe不如Theano。
开发环境搭建：
一、没有GPU
[email protected]-virtual-machine:~$ lspci | grep -i nvidia
[email protected]-virtual-machine:~$ 二、ubuntu版本
[email protected]:~$ uname -m && cat /etc/*release
x86_64
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=15.10
DISTRIB_CODENAME=wily
DISTRIB_DESCRIPTION="Ubuntu 15.10"
NAME="Ubuntu"
VERSION="15.10 (Wily Werewolf)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 15.10"
VERSION_ID="15.10"
HOME_URL="http://www.ubuntu.com/"
SUPPORT_URL="http://help.ubuntu.com/"
BUG_REPORT_URL="http://bugs.launchpad.net/ubuntu/"
[email protected]:~$
三、gcc
[email protected]virtual-machine:~$ gcc --version
gcc.real (Ubuntu 5.2.1-22ubuntu2) 5.2.1 20151010
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
[email protected]virtual-machine:~$
四、安装依赖库
[email protected]-virtual-machine:~$ sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
[sudo] password for learning:
Reading package lists... Done
Building dependency tree 
sudo apt-get install --no-install-recommends libboost-all-dev
sudo apt-get install libatlas-base-dev
五、安装python
六、安装Opencv
安装Opencv
http://blog.csdn.net/forest_world/article/details/51372703Ubuntu
七、安装依赖库
[email protected]:~$ sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
[sudo] password for learning:
Reading package lists… Done
八、下载Caffe
[email protected]:~$ git clone git://github.com/BVLC/caffe.git
Cloning into ‘caffe‘...
remote: Counting objects: 34637, done.
Receiving objects: 100% (34637/34637), 47.81 MiB | 81.00 KiB/s, done.
remote: Total 34637 (delta 0), reused 0 (delta 0), pack-reused 34636
Resolving deltas: 100% (23287/23287), done.
Checking connectivity... done.
九、修改
Makefile 修改：
LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial opencv_core opencv_highgui opencv_imgproc opencv_imgcodecs
这一块代码不需要修改
修改处：
caffe/examples/cpp_classification/classification.cpp文件
十、编译
[email protected]:~/caffe$ cp Makefile.config.example Makefile.config
[email protected]:~/caffe$ ls
caffe.cloc data INSTALL.md matlab src
cmake docker LICENSE models tools
CMakeLists.txt docs Makefile python
CONTRIBUTING.md examples Makefile.config README.md
CONTRIBUTORS.md include Makefile.config.example scripts
[email protected]:~/caffe$ [email protected]:~/caffe$ gedit Makefile.config
[email protected]:~/caffe$ make all
出现问题：
[email protected]:~/caffe$ make all
PROTOC src/caffe/proto/caffe.proto
CXX .build_release/src/caffe/proto/caffe.pb.cc
CXX src/caffe/data_transformer.cpp
CXX src/caffe/common.cpp
CXX src/caffe/internal_thread.cpp
CXX src/caffe/blob.cpp
CXX src/caffe/data_reader.cpp
CXX src/caffe/parallel.cpp
CXX src/caffe/util/hdf5.cpp
In file included from src/caffe/util/hdf5.cpp:1:0:
./include/caffe/util/hdf5.hpp:6:18: fatal error: hdf5.h: No such file or directory
compilation terminated.
Makefile:572: recipe for target ‘.build_release/src/caffe/util/hdf5.o‘ failed
make: *** [.build_release/src/caffe/util/hdf5.o] Error 1
[email protected]:~/caffe$
解决：
Makefile.config
INCLUDE_DIRS
/usr/include/hdf5/serial/
Makefile
LIBRARIES
hdf5_hl and hdf5 改为 hdf5_serial_hl ，hdf5_serial
出现问题：
LD -o .build_release/lib/libcaffe.so.1.0.0-rc3
CXX tools/finetune_net.cpp
CXX/LD -o .build_release/tools/finetune_net.bin
CXX tools/net_speed_benchmark.cpp
CXX/LD -o .build_release/tools/net_speed_benchmark.bin
CXX tools/compute_image_mean.cpp
CXX/LD -o .build_release/tools/compute_image_mean.bin
.build_release/lib/libcaffe.so: undefined reference to `cv::imread(cv::String const&, int)‘
.build_release/lib/libcaffe.so: undefined reference to `cv::imencode(cv::String const&, cv::_InputArray const&, std::vector<unsigned char, std::allocator<unsigned char> >&, std::vector<int, std::allocator<int> > const&)‘
.build_release/lib/libcaffe.so: undefined reference to `cv::imdecode(cv::_InputArray const&, int)‘
collect2: error: ld returned 1 exit status
Makefile:616: recipe for target ‘.build_release/tools/compute_image_mean.bin‘ failed
make: *** [.build_release/tools/compute_image_mean.bin] Error 1
解决方法：
Makefile
修改：
LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial opencv_core opencv_highgui opencv_imgproc opencv_imgcodecs
编译成功：
make test
make runtest
[----------] 2 tests from BatchReindexLayerTest/0, where TypeParam = caffe::CPUDevice<float>
[ RUN ] BatchReindexLayerTest/0.TestForward
[ OK ] BatchReindexLayerTest/0.TestForward (0 ms)
[ RUN ] BatchReindexLayerTest/0.TestGradient
[ OK ] BatchReindexLayerTest/0.TestGradient (373 ms)
[----------] 2 tests from BatchReindexLayerTest/0 (374 ms total)
[----------] Global test environment tear-down
[==========] 1058 tests from 146 test cases ran. (134225 ms total)
[ PASSED ] 1058 tests.
[email protected]virtual-machine:~/caffe$
十一、配置pycaffe
sudo apt-get install python-numpy python-scipy python-matplotlib python-sklearn python-skimage python-h5py python-protobuf python-leveldb python-networkx python-nose python-pandas python-gflags Cython ipython
sudo apt-get install protobuf-c-compiler protobuf-compiler
[email protected]:~/caffe$ make pycaffe
learning@learning-virtual-machine:~/caffe$ make pycaffe
CXX/LD -o python/caffe/_caffe.so python/caffe/_caffe.cpp
touch python/caffe/proto/__init__.py
PROTOC (python) src/caffe/proto/caffe.proto
learning@learning-virtual-machine:~/caffe$
sudo gedit /etc/profile
末尾添加： export PYTHONPATH=/path/to/caffe/python:$PYTHONPATH
用完整路径
source /etc/profile
[email protected]:~/caffe$ python
Python 2.7.10 (default, Oct 14 2015, 16:09:02)
[GCC 5.2.1 20151010] on linux2
Type “help”, “copyright”, “credits” or “license” for more information.
.>>>
出现问题：
.>>> import caffe
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: No module named caffe解决方法：
sudo gedit /etc/profile
export PYTHONPATH=$PYTHONPATH:/home/learning/caffe/python
source /etc/profile
补充:baidu解释
Python（英国发音：/?pa?θ?n/ 美国发音：/?pa?θɑ?n/）, 是一种面向对象、解释型计算机程序设计语言，由Guido van Rossum于1989年发明，第一个公开发行版发行于1991年。
Python是纯粹的自由软件， 源代码和解释器CPython遵循 GPL(GNU General Public License)协议[1] 。
Python语法简洁清晰，特色之一是强制用空白符(white space)作为语句缩进。
Python具有丰富和强大的库。它常被昵称为胶水语言，能够把用其他语言制作的各种模块（尤其是C/C++）很轻松地联结在一起。常见的一种应用情形是，使用Python快速生成程序的原型（有时甚至是程序的最终界面），然后对其中[2] 有特别要求的部分，用更合适的语言改写，比如3D游戏中的图形渲染模块，性能要求特别高，就可以用C/C++重写，而后封装为Python可以调用的扩展类库。需要注意的是在您使用扩展类库时可能需要考虑平台问题，某些可能不提供跨平台的实现。
参考资料：Ubuntu14.04 安装Caffe
http://blog.csdn.net/u011762313/article/details/47262549