Ubuntu14.04(64bit)+caffe+cuda8.0+cudnnV5安装配置（十步走）

电脑配置：ubuntu14.04 64bit (网上有安装双系统的教程，不建议在虚拟机上安装除非只使用cpu版caffe)内存8G显卡GT 755Mcpu i5软件版本: caffe官网当天git下载, nvidia官网下载cuda8.0, cudnnV5Caffe安装步骤如下：一、安装开发所需依赖包sudo apt-get install build-essent

ABadCandy

1881人浏览 · 2017-03-28 10:16:05

ABadCandy · 2017-03-28 10:16:05 发布

电脑配置：

ubuntu14.04 64bit (网上有安装双系统的教程，不建议在虚拟机上安装除非只使用cpu版caffe)

内存8G

显卡GT 755M

cpu i5

软件版本: caffe官网当天git下载, nvidia官网下载cuda8.0, cudnnV5

Caffe安装步骤如下：

一、安装开发所需依赖包

sudo apt-get install build-essential # basic requirement

sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libboost-all-dev libhdf5-serial-dev libgflags-dev libgoogle-glog-dev liblmdb-dev protobuf-compiler #required by caffe

二、安装Cuda8.0

1. 离线.deb安装：官网下载地址: https://developer.nvidia.com/cuda-downloads

安装之前请先进行md5校验，确保下载的安装包完整

2. 切换到下载的deb所在目录，执行下边的命令

sudo dpkg -i cuda-repo-ubuntu1404-8-0-local_8.0-18_amd64.deb

sudo apt-get update

sudo apt-get install cuda

3. 设置cuda的环境变量

（1）使用如下命令：

sudo vim /etc/profile

在/etc/profile中添加CUDA环境变量末尾添加如下内容：

PATH=/usr/local/cuda/bin:$PATH

export PATH

（2）保存后, 执行下列命令, 使环境变量立即生效

source /etc/profile

（3）同时需要添加lib库路径：使用

sudo vim/etc/ld.so.conf.d/cuda.conf

在/etc/ld.so.conf.d/加入cuda.conf文件, 内容如下

/usr/local/cuda/lib64

（4）保存后，执行下列命令使之立刻生效

sudo ldconfig

（5）然后重启电脑

sudo reboot

三、安装CudnnV5(最新的为6.0版本)

1. 官网上注册申请下载cudnn-8.0-linux-x64-v5.0-ga.tgz（地址https://developer.nvidia.com/cudnn)

2. 到下载目录下执行如下命令：

$ tar-zxvf cudnn-8.0-linux-x64-v5.0-ga.tgz

$ cd cuda/include

$ sudo cp *.h /usr/local/cuda/include/

$ cd ../lib64

$ sudo cp lib* /usr/local/cuda/lib64/

$ cd /usr/local/cuda/lib64/

$ sudo chmod +r libcudnn.so.5.0.5

$ sudo ln -sf libcudnn.so.5.0.5 libcudnn.so.5

$ sudo ln -sf libcudnn.so.5 libcudnn.so

$ sudo ldconfig

四、安装Cuda Sample（验证）

1. 进入/usr/local/cuda/samples,执行下列命令来编译samples：

$ cd /usr/local/cuda/samples

$ sudo make all -j4（依自己电脑核数而定）

2. 全部编译完成后，进入/usr/local/cuda/samples/bin/x86_64/linux/release, 运行deviceQuery

$ cd /usr/local/cuda/samples/bin/x86_64/linux/release

$ ./deviceQuery

如果出现类似如下显卡信息，则驱动及显卡安装成功：

./deviceQuery Starting...

CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "GeForce GTX 670"

CUDA Driver Version / Runtime Version 8.0 / 8.0

CUDA Capability Major/Minor version number: 3.0

Total amount of global memory: 4095 MBytes (4294246400 bytes)

( 7) Multiprocessors, (192) CUDA Cores/MP: 1344 CUDA Cores

GPU Clock rate: 1098 MHz (1.10 GHz)

Memory Clock rate: 3105 Mhz

Memory Bus Width: 256-bit

L2 Cache Size: 524288 bytes

Maximum Texture Dimension Size (x,y,z) 1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)

Maximum Layered 1D Texture Size, (num) layers 1D=(16384), 2048 layers

Maximum Layered 2D Texture Size, (num) layers 2D=(16384, 16384), 2048 layers

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 49152 bytes

Total number of registers available per block: 65536

Warp size: 32

Maximum number of threads per multiprocessor: 2048

Maximum number of threads per block: 1024

Max dimension size of a thread block (x,y,z): (1024, 1024, 64)

Max dimension size of a grid size (x,y,z): (2147483647, 65535, 65535)

Maximum memory pitch: 2147483647 bytes

Texture alignment: 512 bytes

Concurrent copy and kernel execution: Yes with 1 copy engine(s)

Run time limit on kernels: Yes

Integrated GPU sharing Host Memory: No

Support host page-locked memory mapping: Yes

Alignment requirement for Surfaces: Yes

Device has ECC support: Disabled

Device supports Unified Addressing (UVA): Yes

Device PCI Bus ID / PCI location ID: 1 / 0

Compute Mode:

< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 8.0, CUDA Runtime Version = 8.0, NumDevs = 1, Device0 = GeForce GTX 670

Result = PASS

NOTE:上边的显卡信息是从别的地方拷过来的，我的GT755M显卡不是这些信息，如果没有这些信息，那肯定是安装不成功，找原因吧！

五、安装Inter MKL或openblas或Atlas(矩阵加速库)

装MKL需要上Intel官网用edu邮箱申请，在这我装的Atlas，安装命令：

$ sudo apt-get install libatlas-base-dev

六、安装OpenCV

1. 下载并编译OpenCV（官网原版OpenCV：http://opencv.org/），或者使用修改版的安装包(百度云盘链接: http://pan.baidu.com/s/1qX1uFHa 密码:wysa)（下面的安装方式使用该包完成，安装包修改了dependencies.sh文件并增加了OpenCV 3.0.0的安装文件）

2. 切换到文件保存的文件夹，然后安装依赖项：(进入 “Install-OpenCV-master” 文件夹下)

$ sudo sh Ubuntu/dependencies.sh

3. 切换目录Ubuntu/3.0/安装OpenCV 3.0.0：

$ cd Ubuntu/3.0/

$ sudo sh opencv3_0_0.sh

保证网络畅通，因为软件需要联网这里时间较长，请耐心等待。

七、安装Pycaffe环境

1. 按caffe官网的推荐使用Anaconda，去Anaconda官网下安装包(下载python2.7 的Anaconda2-4.0.0-Linux-x86_64.sh)

2. 切换到文件所在目录，执行

bash Anaconda2-4.0.0-Linux-x86_64.sh

NOTE:后边的文件名按自己下的版本号更改，整个安装过程请选择默认,安装路径为 /home/username/anaconda(username为自己系统的用户名)

3. 添加Anaconda Library Path：使用

$ sudo vim /etc/ld.so.conf

在/etc/ld.so.conf最后加入以下路径 (NOTE:下边的username要替换为自己的用户名)：

/home/username/anaconda/lib

4. 使用

vim ~/.bashrc

在~/.bashrc最后添加下边路径：

export LD_LIBRARY_PATH="/home/username/anaconda/lib:$LD_LIBRARY_PATH"

5. 退出运行如下命令使配置文件生效:

$ sudo ldconfig

八、安装python依赖库

1. 去caffe的github下载caffe源码包（https://github.com/BVLC/caffe ）

2. 进入caffe-master下的python目录，执行如下命令

sudo apt-get install python-pip #如果没有安装pip则需要执行该命令

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

九、编译Caffe

1. 进入caffe-master目录，复制一份Makefile.config.examples：

cp Makefile.config.example Makefile.config

2. 修改配置文件，如果前边和该安装教程一致，那么配置文件应该是这个样子的：

## Refer tohttp://caffe.berkeleyvision.org/installation.html

# Contributions simplifying and improving ourbuild system are welcome!

# cuDNN acceleration switch (uncomment to buildwith cuDNN).

USE_CUDNN := 1

# CPU-only switch (uncomment to build without GPUsupport).

# CPU_ONLY := 1

# uncomment to disable IO dependencies andcorresponding data layers

# USE_OPENCV := 0

# USE_LEVELDB := 0

# USE_LMDB := 0

# uncomment to allow MDB_NOLOCK when reading LMDBfiles (only if necessary)

# Youshould not set this flag if you will be reading LMDBs with any

# possibility of simultaneous read and write

# ALLOW_LMDB_NOLOCK := 1

# Uncomment if you're using OpenCV 3

OPENCV_VERSION := 3

# To customize your choice of compiler, uncommentand set the following.

# N.B. the default for Linux is g++ and thedefault for OSX is clang++

# CUSTOM_CXX := g++

# CUDA directory contains bin/ and lib/directories that we need.

CUDA_DIR := /usr/local/cuda

# On Ubuntu 14.04, if cuda tools are installedvia

# "sudo apt-get installnvidia-cuda-toolkit" then use this instead:

# CUDA_DIR := /usr

# CUDA architecture setting: going with all ofthem.

# For CUDA < 6.0, comment the *_50 lines forcompatibility.

CUDA_ARCH := -gencode arch=compute_20,code=sm_20\

-gencode arch=compute_20,code=sm_21 \

-gencode arch=compute_30,code=sm_30 \

-gencode arch=compute_35,code=sm_35 \

-gencode arch=compute_50,code=sm_50 \

-gencode arch=compute_50,code=compute_50

# BLAS choice:

# atlas for ATLAS (default)

# mkl for MKL

# open for OpenBlas

BLAS := atlas #这里根据自己安装的加速库而定

# NOTE: this is required only if you will compilethe python interface.

# We need to be able to find Python.h andnumpy/arrayobject.h.

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

/usr/lib/python2.7/dist-packages/numpy/core/include

# Anaconda Python distribution is quite popular.Include path:

# Verify anaconda location, sometimes it's inroot.

ANACONDA_HOME := /home/username/anaconda #记得改username

PYTHON_INCLUDE :=$(ANACONDA_HOME)/include \

$(ANACONDA_HOME)/include/python2.7 \

$(ANACONDA_HOME)/lib/python2.7/site-packages/numpy/core/include \

# Uncomment to use Python 3 (default is Python 2)

# PYTHON_LIBRARIES := boost_python3 python3.5m

# PYTHON_INCLUDE := /usr/include/python3.5m \

# /usr/lib/python3.5/dist-packages/numpy/core/include

# We need to be able to find libpythonX.X.so or.dylib.

#PYTHON_LIB := /usr/lib

PYTHON_LIB := $(ANACONDA_HOME)/lib

# Homebrew installs numpy in a non standard path(keg only)

# PYTHON_INCLUDE += $(dir $(shell python -c'import numpy.core; print(numpy.core.__file__)'))/include

# PYTHON_LIB += $(shell brew --prefix numpy)/lib

# Uncomment to support layers written in Python(will link against Python libs)

WITH_PYTHON_LAYER := 1

# Whatever else you find you need goes here.

INCLUDE_DIRS := $(PYTHON_INCLUDE)/usr/local/include /usr/lib/x86_64-linux-gnu/hdf5/serial/include

LIBRARY_DIRS := $(PYTHON_LIB)/usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu/hdf5/serial

# If Homebrew is installed at a non standardlocation (for example your home directory) and you use it for generaldependencies

# INCLUDE_DIRS += $(shell brew --prefix)/include

# LIBRARY_DIRS += $(shell brew –prefix)/lib

# Uncomment to use `pkg-config` to specify OpenCVlibrary paths.

# (Usually not necessary -- OpenCV libraries arenormally installed in one of the above $LIBRARY_DIRS.)

# USE_PKG_CONFIG := 1

# N.B. both build and distribute dirs are clearedon `make clean`

BUILD_DIR := build

DISTRIBUTE_DIR := distribute

# Uncomment for debugging. Does not work on OSXdue to https://github.com/BVLC/caffe/issues/171

# DEBUG := 1

# The ID of the GPU that 'make runtest' will useto run unit tests.

TEST_GPUID := 0

# enable pretty build (comment to see fullcommands)

Q ?= @

3. 保存退出，然后编译：(如报错可按报错内容安装相应依赖包或修改环境变量)

$ make all -j4

$ make test

$ make runtest

4. 编译python wrapper

$ make pycaffe

$ vim ~/.bashrc 在末尾添加：

export PYTHONPATH=”/home/username/caffe-master/python:$PYTHONPATH”

$ sudo ldconfig #使之生效

十、使用Mnist数据集测试

Caffe默认情况会安装在$CAFFE_ROOT，就是解压到那个目录，例如：$home/username/caffe-master，所以下面的工作，默认已经切换到了该工作目录。下面的工作主要是，用于测试Caffe是否工作正常，不做详细评估。具体设置请参考官网：http://caffe.berkeleyvision.org/gathered/examples/mnist.html

1. 数据预处理