ubuntu 16.04 安装caffe 详细，全面

install caffe on ubuntu 16.04
可以参看视频：youtube http://caffe.berkeleyvision.org

1.官网的安装说明

documentation --> installation instrument

Step-by-step Instructions: ->Ubuntu installation the standard platform

2.接下来就是在命令窗口复制、粘贴、执行命令了
For Ubuntu (< 17.04) . #分2条执行

注解：黄色标记的文本就是要执行的命令

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

CUDA: Install by apt-get or the NVIDIA
This can be skipped for CPU-only installation.

BLAS: install ATLAS by sudo apt-get install libatlas-base-dev or install OpenBLAS by sudo apt-get install libopenblas-dev or MKL for better CPU performance.

Python (optional): if you use the default Python you will need to sudo apt-get install the python-dev package to have the Python headers for building the pycaffe interface.
sudo apt-get install python

Compatibility notes, 16.04
CUDA 8 is required on Ubuntu 16.04.
install cuda8 #视频没有做，建议提前安装上

Remaining dependencies, 14.04
Everything is packaged in 14.04.
sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

Continue with compilation.

Python
The main requirements are numpy and boost.python (provided by boost). pandas is useful too and needed for some examples.
#以下是预备工作
require https://github.com/BVLC/caffe

https://github.com/BVLC/caffe.git #复制这行

mkdir .local/install #创建文件夹
cd .local/install #进入文件夹
git clone https://github.com/BVLC/caffe.git #下载到本地文件夹，
sudo apt install git #第一次安装出现

补充：
.local/install 的具体位置 ：  /home/~/.local/install
~/ #指的是用户的文件夹
.文件为隐藏的文件  #查看方法：进入自己主目录，按ctrl+h.就能看见以点号开头的隐藏文件

 ls -a  显示所有文件，也包括隐藏文件
删除文件夹 命令：sudo rm -r 文件夹名

cd caffe #不换路径，直接cd即可
ls #查看当前路径下的文件
find -name requirements.txt #找到requirements

cd python #进入此文件夹，即可找到requirement.txt 文件
sudo apt install python-pip #第一次安装出现
==for req in $(cat requirements.txt); do pip install $req; done #安装requirement里的依赖包 ==

cd … #回到caffe文件下

but we suggest first installing the Anaconda Python distribution, which provides most of the necessary packages, as well as the hdf5 library dependency. #建议安装anconda

python -m site #查看python目录，找到以site-package结尾的，copy路径

3.Compilation 编译
Caffe can be compiled with either Make or CMake. Make is officially supported while CMake is supported by the community.

ls #目的是下面的命令输入时，可以参考文件名
cp Makefile.config.example Makefile.config #复制文件
vi Makefile.config #此处的编辑可以用 xcode

#修改的makefile.config的地方有两处，如下：
CPU_ONLY := 1 #去掉注释

==PYTHON_INCLUDE := /usr/include/python2.7 ==
#在此处粘贴刚才复制的site-package的路径，替换掉下面的路径
/usr/lib/python2.7/dist-packages/numpy/core/include #该行不要

：wq #退出并保存
ls /。。。/sitepackage前的路径
find / …/site-packages -name numpy

vi Makefile.config #注意请提前切换路径
#在此处粘贴刚才复制的site-package的路径 添加 numpy/core/include

find /Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/ -name numpy #
#执行完上条命令，可以看到有 numpy/core/include/numpy
/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages//numpy/core/include/numpy

==WITH_PYTHON_LAYER := 1 ==#该行的注释去掉

添加
/usr/lib/x86_64-linux-gnu/hdf5

查看linux下的opencv安装路径：
sudo find / -iname “opencv”

4.遇到的问题
编译 pycaffe时报错：fatal error: numpy/arrayobject.h没有那个文件或目录
sudo apt-get install python-numpy
然后
sudo make pycaffe -j16

wget http://archive.ipython.org/release/5.1.0/ipython-5.1.0.tar.gz
tar -xzf ipython-5.1.0.tar.gz
cd ipython-5.1.0
sudo python setup.py install

ImportError: cannot import name '_validate_lengths
numpy版本太高，我对numpy进行了重新安装：
pip install numpy1.15.0==

ImportError: No module named scipy
pip install 'scipy0.13.2’==

cannot import name _denoise_bilateral
pip install 'scikit-image0.9.3’==

ImportError: No module named pydot
sudo apt-get install graphviz
pip install pydot

export PYTHONPATH=/home/cody/.local/install/caffe/python

/home/cody/.local/lib/python2.7/site-packages/scipy/stats/stats.py:252: RuntimeWarning: numpy.ufunc size changed, may indicate binary incompatibility
from ._rank import rankdata, tiecorrect
解决办法：
pip uninstall numpy 1.15.0
pip install numpy1.16.4

以上，如有问题，请百度或这Google或者GitHub，对，GitHub才是真正解决问题的地方