1. apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libboost-all-dev libhdf5-dev libgflags-dev libgoogle-glog-dev liblmdb-dev protobuf-compiler

2. git clone https://github.com/BVLC/caffe

cd caffe
cp Makefile.config.example Makefile.config

3. pip install scikit-image protobuf

  cd python for req in $(cat requirements.txt); do sudo pip install $req; done

If TypeError: unsupported operand type(s) for -=: 'Retry' and 'int',

Reason: The computer connects with internet through proxy, so add the --proxy parameter like this:

sudo pip install --proxy=http://dev-myproxy.com:8080 scikit-image protobuf

4. modify the Makefile.config. Uncomment the line CPU_ONLY := 1, and the line OPENCV_VERSION := 3 if only use cpu and opencv3.

5. make all

CXX src/caffe/net.cpp

src/caffe/net.cpp:8:18: fatal error: hdf5.h: No such file or directory compilation terminated. Makefile:575: recipe for target '.build_release/src/caffe/net.o' failed make: *** [.build_release/src/caffe/net.o] Error 1

 

This is because the headfiles of hdf5 is in /usr/include/hdf5/serial/, copy all of them into /usr/include directory.

And then

sudo ln -s /usr/lib/x86_64-linux-gnu/libhdf5_serial.so.10.1.0 /usr/lib/x86_64-linux-gnu/libhdf5.so

sudo ln -s /usr/lib/x86_64-linux-gnu/libhdf5_serial_hl.so.10.0.2 /usr/lib/x86_64-linux-gnu/libhdf5_hl.so

 

6. make all again, there will be another error: cann't find cblas.h, no such file or directory.

View the Makefile.config, the option BLAS is atlas, so just install libatlas-base-dev:

apt-get install libatlas-base-dev

7. ld error: can't find -lopencv_imgdecodecs

Wrong opencv version will lead this error.

dpk-config --modversion opencv #check openvcv's version

and then modify the OPENCV_VERSION option in Makefile.config accordingly.

8. make test

9. make runtest

10. make pycaffe

python/caffe/_caffe.cpp:10:31: fatal error: numpy/arrayobject.h: No such file or directory

It was caused because numpy was installed in a different path, and we must manually point to it.

 vi Makefile.config from this:

PYTHON_INCLUDE := /usr/include/python2.7 \

/usr/lib/python2.7/dist_packages/numpy/core/include

to:

PYTHON_INCLUDE :=/usr/include/python2.7 \

/usr/local/lib/python2.7/dist-package/numpy/core/inlcude

 11. Modify PYTHONPATH environment parameter

vi .bashrc

export PYTHONPATH=xxx/caffe/python:$PYTHONPATH

Then we can import caffe in python like this:

import caffe

references:

https://chunml.github.io/ChunML.github.io/project/Installing-Caffe-CPU-Only/

Compile caffe on unbutu 16.0.4的更多相关文章

  1. Ubuntu16.04+Cuda8.0+1080ti+caffe+免OpenCV3.2.0+faster-rCNN教程

    一.事先声明:1.Ubuntu版本:Ubuntu使用的是16.04.而不是16.04.1或16.04.2,这三个是有区别的.笔者曾有过这样的经历,Git上一个SLAM地图构建程序在Ubuntu14.0 ...

  2. ubuntu 16.04源码编译和配置caffe详细教程 | Install and Configure Caffe on ubuntu 16.04

    本文首发于个人博客https://kezunlin.me/post/b90033a9/,欢迎阅读! Install and Configure Caffe on ubuntu 16.04 Series ...

  3. Caffe学习系列(16):caffemodel可视化

    通过前面的学习,我们已经能够正常训练各种数据了.设置好solver.prototxt后,我们可以把训练好的模型保存起来,如lenet_iter_10000.caffemodel. 训练多少次就自动保存 ...

  4. [boost] build boost with intel compiler 16.0.XXX

    Introduction There are few information about how to compile boost with Intel compiler. This article ...

  5. ubuntu 16.04源码编译OpenCV教程 | compile opencv on ubuntu 16.04

    本文首发于个人博客https://kezunlin.me/post/15f5c3e8/,欢迎阅读! compile opencv on ubuntu 16.04 Series Part 1: comp ...

  6. ubuntu 16.04上源码编译dlib教程 | compile dlib on ubuntu 16.04

    本文首发于个人博客https://kezunlin.me/post/c6ead512/,欢迎阅读! compile dlib on ubuntu 16.04 Series Part 1: compil ...

  7. ubuntu 16.04上源码编译opengv | compile opengv on ubuntu 16.04

    本文首发于个人博客https://kezunlin.me/post/1e5d14ee/,欢迎阅读! compile opengv on ubuntu 16.04 Series compile open ...

  8. 安装MYSQL详细教程 版本:mysql-installer-community-5.7.16.0 免安装版本和安装版本出现错误的解决

    一.版本的选择 之前安装的Mysql,现在才来总结,好像有点晚,后台换系统了,现在从新装上Mysql,感觉好多坑,我是来踩坑,大家看到坑就别跳了,这样可以省点安装时间,这个折腾了两天,安装了好多个版本 ...

  9. Navicat Premium 12.1.16.0安装与激活

    声明:本文所提供的所有软件均来自于互联网,仅供个人研究和学习使用,请勿用于商业用途,下载后请于24小时内删除,请支持正版! 本文介绍Navicat Premium 12的安装.激活与基本使用.已于20 ...

随机推荐

  1. [转]numpy 100道练习题

    100 numpy exercise 翻译:YingJoy 网址: https://www.yingjoy.cn/ 来源:https://github.com/rougier/numpy-100 Nu ...

  2. TCP连接数过多问题

        在一次生产上线后,发现使用的 8086 端口相关的 TCP 连接数竟然多大 6K+ ,有时候甚至会逼近 1w ,这个数量对于一个只是在内部使用的监控系统来说, 无论如何都是无法接受的, 于是开 ...

  3. Xcode真机调试失败:The identity used to sign the executable is no longer valid

    在Xcode中突然好久没有使用真机调试了.今天使用真机的时候.出现例如以下的警告.并真机执行失败: The identity used to sign the executable is no lon ...

  4. 通过一个小Trick实现shader的像素识别/统计操作

    2018/12/14日补充:后来发现compute shader里用AppendStructuredBuffer可以解决这类问题,请看这里:https://www.cnblogs.com/hont/p ...

  5. atitit r9 doc on home ntpc .docx

    卷 p2soft 的文件夹 PATH 列表 卷序列号为 9AD0-D3C8 D:. │  Aittit pato 面对拒绝  的回应.docx │  Atitit  中国明星数量统计 attilax. ...

  6. jQuery创建元素和添加子元素

    var li = $("<li class=\"TopNav arrow\" secondMenu=\"menu_" + i + "\ ...

  7. RSA 分段加解密【解决“不正确的长度”的异常】

    RSA 是常用的非对称加密算法.最近使用时却出现了“不正确的长度”的异常,研究发现是由于待加密的数据超长所致. .NET Framework 中提供的 RSA 算法规定: 待加密的字节数不能超过密钥的 ...

  8. input 呼起数字键盘

    1. Android,定义 type="number" 2. iOS,定义 style="ime-mode: disabled;" 注,呼起数字键盘后,用户输入 ...

  9. datatable to List<T>带缓存

    public class DataHelper { //datarow 转换的类型缓存 private static MemoryCache modelCash = MemoryCache.Defau ...

  10. python数据分析及展示(二)

    一.Matplotlib库入门 1. Matplotlib库的介绍 http://matplotlib.org/gallery.html可查看Matplotlib库的效果 Matplotlib库由各种 ...