从零到放弃之配置faster_rcnn所有过程

配置：

NVIDIA x86-64-390.48.run
CUDA9.0.176_384.81_linux.run
CUDNN9.0-linux-x64-v7.1.tgz

1.Nvidia 驱动正确安装过程

卸载掉原有驱动

sudo apt-get remove –purge nvidia*

如果使用的是 apt-get 安装可以使用这种方法卸载，如果使用的是 runfile，则使用--uninstall 命令，当然 runfile 安装的时候会卸载掉之前的驱动，所以可以不用手动去卸载。

禁用 nouveau

打开编辑配置文件：

sudo gedit /etc/modprobe.d/blacklist.conf

在最后一行添加：

blacklist nouveau

禁用 nouveau 第三方驱动，之后也不需要改回来

执行：

sudo update-initramfs -u

重启后执行：

lsmod | grep nouveau

没有输出即屏蔽好了

安装驱动

进入命令行界面注：（Ctrl + Alt + F7 是回到桌面系统）

Ctrl-Alt+F1

给驱动 run 文件赋予执行权限

sudo chmod a+x NVIDIA-Linux-x86_64-390.48.run //获取权限

禁用 X（显卡）服务（至关重要）

执行：

sudo /etc/init.d/lightdm stop

安装(注意参数)

sudo ./NVIDIA-Linux-x86_64-390.48.run --no-opengl-files

#注意，这里总是显示无效指令–no-opengl-files ，自己手打一遍就好

sudo /etc/init.d/lightdm start #开启显卡

–no-opengl-files 只安装驱动文件，不安装 OpenGL 文件。这个参数最重要
–no-x-check 安装驱动时不检查 X 服务
–no-nouveau-check 安装驱动时不检查 nouveau

后面两个参数可不加。

安装驱动的时候，有一布问你”Would you like to run the nvidia-xconfig utility to automatically update your X configuration file…”什么的，选择 No。

重启电脑，没有问题，输入命令：

nvidia-smi

2.安装 cuda（CUDA 是一种并行计算的模型，能利用英伟达 GPU 的并行计算引擎）

下载好 CUDA Toolkit9.1 后，执行如下代码进行安装（此处不需要安装 OPGL），代码如下：

1 sudo sh cuda_9.0.176_384.81_linux.run --no-opengl-libs #run 文件的文件名根据自己下的文件名修改，默认是我提供的文件

输出显示：

这里先进入协议，可以按回车阅读，也可以直接 CTRL+C 退出

Do you accept the previously read EULA?
accept/decline/quit: accept
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 384.81?
(y)es/(n)o/(q)uit: n
Install the CUDA 9.0 Toolkit?
(y)es/(n)o/(q)uit: y
Enter Toolkit Location
[ default is /usr/local/cuda-9.0 ]:
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: y
Install the CUDA 9.0 Samples?
(y)es/(n)o/(q)uit: y 13 Enter CUDA Samples Location 14 [ default is /home/pertor ]:

Installing the CUDA Toolkit in /usr/local/cuda-9.0 ...
Missing recommended library: libXmu.so

添加环境变量：

sudo gedit ~/.bashrc

export PATH=/usr/local/cuda-9.0/bin${PATH:+:${PATH}}

export LD_LIBRARY_PATH=/usr/local/cuda9.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

source ~/.bashrc

验证 CUDA9.0 是否安装成功

cd /usr/local/cuda-9.0/samples/1_Utilities/deviceQuery

sudo make

./deviceQuery

输出如下信息表示成功安装

./deviceQuery Starting...

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

Detected 1 CUDA Capable device(s)

Device 0: "GeForce GT 740M"

CUDA Driver Version / Runtime Version 9.0 / 9.0

CUDA Capability Major/Minor version number: 3.5

Total amount of global memory: 2004 MBytes (2100953088 bytes)

( 2) Multiprocessors, (192) CUDA Cores/MP: 384 CUDA Cores

GPU Max Clock rate: 1032 MHz (1.03 GHz)

Memory Clock rate: 800 Mhz

Memory Bus Width: 64-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

3.安装 cudnn(NVIDIA 打造的针对深度神经网络的加速库)

执行如下步骤：

tar -zxvf cudnn-9.0-linux-x64-v7.1.tgz
sudo cp cuda/include/cudnn.h /usr/local/cuda/include/
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/ -d
sudo chmod a+r /usr/local/cuda/include/cudnn.h
sudo chmod a+r /usr/local/cuda/lib64/libcudnn*

复制动态链接库

cd /usr/local/cuda/lib64/ sudo rm -rf libcudnn.so libcudnn.so.7

生成软衔接

sudo ln -s libcudnn.so.7.1.2 libcudnn.so.7

生成软链接

sudo ln -s libcudnn.so.7libcudnn.so

执行完如上命令之后，cuDNN 就安装好了，这时我们可以发现在 /usr/local/cuda/include 目录下就多了 cudnn.h 头文件。

终端中执行 nvcc -V 显示如下信息则表示成功

nvcc -V

pertor@pertor-computer:~$ nvcc -V nvcc: NVIDIA (R) Cuda compiler driver

Built on Fri_Sep__1_21:08:03_CDT_2017

Cuda compilation tools, release 9.0, V9.0.176

4.安装 opencv3.4.1

安装之前需要安装 cmake

安装 cmake: sudo apt-get install cmake

1.下载 opencv

下载地址：https://github.com/opencv/opencv

2.下载之后解压安装包

解压：unzip opencv-3.4.1.zip

3.在当前目录下创建一个新文件夹用于编译 opencv

创建文件夹：mkdir opencv3.4_build

4.进入刚刚创建的文件夹里面

进入文件夹：cd opencv_build

5.编译 opencv ,进行 cmake

cmake ../opencv-3.4.1 -DWITH_GTK_2_X=ON -DCMAKE_INSTALL_PREFIX=/usr/local -DCMAKE_BUILD_TYPE=RELEASE -

DCMAKE_INSTALL_PREFIX=/usr/local -DWITH_TBB=ON -DBUILD_NEW_PYTHON_SUPPORT=ON -DWITH_V4L=ON -DINSTALL_C_EXAMPLES=ON -DINSTALL_PYTHON_EXAMPLES=ON -DBUILD_EXAMPLES=ON -DWITH_OPENGL=ON -DENABLE_FAST_MATH=1 -DCUDA_FAST_MATH=1 DWITH_CUBLAS=1 -DWITH_OPENMP=ON

cmake 成功之后进行编译

编译 opencv：make -j8

安装 opencv

安装：sudo make install

6.配置 FASTER-RCNN Caffe

安装相关依赖

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

sudo apt-get install libopenblas-dev liblapack-dev libatlas-base-dev

sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

pip install cython

pip install easydict

pip install python-opencv

1）拉取 faster rcnn 代码

git clone --recursive https://github.com/rbgirshick/py-faster-rcnn.git

2）cd 到 lib 目录，生成 cython

cd py-faster-rcnn/lib
make -j8

3）修改 Makefile.config 文件

cp Makefile.config.example Makefile.config
sudo gedit Makefile.config

进行 Makefile 和 Makefile.config 的修改

内容附上：

Makefile.config：：

## Refer to http://caffe.berkeleyvision.org/installation.html # Contributions simplifying and improving our build system are welcome! # cuDNN acceleration switch (uncomment to build with cuDNN).

USE_CUDNN:=1

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

# CPU_ONLY := 1

# uncomment to disable IO dependencies and corresponding data layers

# USE_OPENCV := 0

# USE_LEVELDB := 0

# USE_LMDB := 0

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

# You should 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, uncomment and set the following.

# N.B. the default for Linux is g++ and the default 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 installed via

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

# CUDA_DIR := /usr

# CUDA architecture setting: going with all of them.

# For CUDA < 6.0, comment the *_50 through *_61 lines for compatibility.

# For CUDA < 8.0, comment the *_60 and *_61 lines for compatibility.

# For CUDA >= 9.0, comment the *_20 and *_21 lines for compatibility.

CUDA_ARCH := -gencode arch=compute_30,code=sm_30 \

-gencode arch=compute_30,code=sm_30 \

-gencode arch=compute_35,code=sm_35 \

-gencode arch=compute_50,code=sm_50 \

-gencode arch=compute_52,code=sm_52 \

-gencode arch=compute_60,code=sm_60 \

-gencode arch=compute_61,code=sm_61 \ -gencode arch=compute_61,code=compute_61

# BLAS choice:

# atlas for ATLAS (default)

# mkl for MKL

# open for OpenBlas

BLAS := atlas

# Custom (MKL/ATLAS/OpenBLAS) include and lib directories. # Leave commented to accept the defaults for your choice of BLAS # (which should work)!

# BLAS_INCLUDE := /path/to/your/blas

# BLAS_LIB := /path/to/your/blas

# Homebrew puts openblas in a directory that is not on the standard search path

# BLAS_INCLUDE := $(shell brew --prefix openblas)/include

# BLAS_LIB := $(shell brew --prefix openblas)/lib

# This is required only if you will compile the matlab interface.

# MATLAB directory should contain the mex binary in /bin.

# MATLAB_DIR := /usr/local

# MATLAB_DIR := /Applications/MATLAB_R2012b.app

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

# We need to be able to find Python.h and numpy/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 in root.

#ANACONDA_HOME := $(HOME)/anaconda2

#PYTHON_INCLUDE := $(ANACONDA_HOME)/include \

# $(ANACONDA_HOME)/include/python2.7 \

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

LIBRARIES += glog gflags protobuf leveldb snappy \ lmdb boost_system hdf5_h1 hdf5 m\ opencv_core opencv_highhui opencv_imgproc opencv_imgcodes opencv_videoio

# 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/include/hdf5/serial

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

# If Homebrew is installed at a non standard location (for example your home directory) and you use it for general dependencies

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

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

# NCCL acceleration switch (uncomment to build with NCCL)

# https://github.com/NVIDIA/nccl (last tested version: v1.2.3-1+cuda8.0)

# USE_NCCL := 1

# Uncomment to use `pkg-config` to specify OpenCV library paths.

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

# USE_PKG_CONFIG := 1

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

BUILD_DIR := build

DISTRIBUTE_DIR := distribute

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

# DEBUG := 1

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

TEST_GPUID := 0

# enable pretty build (comment to see full commands)

Q ?= @

（4） Makefile:

修改 makefile 文件打开 makefile 文件，做如下修改：将：

NVCCFLAGS +=-ccbin=$(CXX)-Xcompiler-fPIC $(COMMON_FLAGS)

替换为：

NVCCFLAGS += -D_FORCE_INLINES -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)

(5)编辑/usr/local/cuda/include/host_config.h

将其中的第 115 行注释掉：将

# error-- unsupported GNU version! gcc versions later than 4.9 are not supported!

改为

//#error-- unsupported GNU version! gcc versions later than 4.9 are not supported!

（6)替换 cudnn 文件

1).将/py-faster-rcnn/caffe-fast-rcnn/include/caffe/util/cudnn.hpp 换成最新版的 caffe 里的 cudnn 的实现，即相应的 cudnn.hpp.

2).将/py-faster-rcnn/caffe-fast-rcnn/include/caffe/layer 里的，所有以 cudnn 开头的文件都替换成最新版的 caffe 里的相应的同名文件

3）将/py-faster-rcnn/caffe-fast-rcnn/src/caffe/layer 里的，所有以 cudnn 开头的文件都替换成最新版的 caffe 里的相应文件

（7）编译 pycaffe

cd ~/caffe
make all -j8
make test -j8
make runtest -j8
make pycaffe

BUG 汇总：

编译 opencv 时碰到如下是错误:

driver_api_multi.cpp:(.text._ZNK6WorkerclEi+0x893): undefined reference to `cuCtxPopCurrent_v2' CMakeFiles/example_gpu_driver_api_multi.dir/driver_api_multi.cpp.o: In function `main': driver_api_multi.cpp:(.text.startup.main+0x11c): undefined reference to `cuInit' driver_api_multi.cpp:(.text.startup.main+0x133): undefined reference to `cuDeviceGet' driver_api_multi.cpp:(.text.startup.main+0x14d): undefined reference to `cuCtxCreate_v2' driver_api_multi.cpp:(.text.startup.main+0x15d): undefined reference to `cuCtxPopCurrent_v2' driver_api_multi.cpp:(.text.startup.main+0x172): undefined reference to `cuDeviceGet' driver_api_multi.cpp:(.text.startup.main+0x18c): undefined reference to `cuCtxCreate_v2' driver_api_multi.cpp:(.text.startup.main+0x19c): undefined reference to `cuCtxPopCurrent_v2' collect2: error: ld returned 1 exit status samples/gpu/CMakeFiles/example_gpu_driver_api_multi.dir/build.make:125: recipe for target 'bin/example_gpu_driver_api_multi' failed make[2]: *** [bin/example_gpu_driver_api_multi] Error 1

CMakeFiles/Makefile2:39358: recipe for target 'samples/gpu/CMakeFiles/example_gpu_driver_api_multi.dir/all' failed make[1]: *** [samples/gpu/CMakeFiles/example_gpu_driver_api_multi.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs....

原因是找不到 cuda 库，解决方式如下(红色为新添加)

/opencv3.41/samples/gpu/CMakeLists.txt

（大概54行左右）

foreach(sample_filename ${all_samples}) ocv_define_sample(tgt ${sample_filename} gpu) ocv_target_link_libraries(${tgt} ${OPENCV_LINKER_LIBS}${OPENCV_CUDA_SAMPLES_REQUIRED_DEPS}) if(HAVE_CUDA AND NOT ANDROID) ocv_target_link_libraries(${tgt} ${CUDA_CUDA_LIBRARY})

endif() if(HAVE_opencv_xfeatures2d) ocv_target_link_libraries(${tgt} opencv_xfeatures2d) endif()

Error 2 :

src/caffe/test/test_smooth_L1_loss_layer.cpp:11:35: fatal error: caffe/vision_layers.hpp: No such file or directory compilation terminated.

Makefile:563: recipe for target '.build_release/src/caffe/test/test_smooth_L1_loss_layer.o' failed make: *** [.build_release/src/caffe/test/test_smooth_L1_loss_layer.o] Error 1 make: *** Waiting for unfinished jobs....

src/caffe/test/test_smooth_L1_loss_layer.cpp:11:35: fatal error: caffe/vision_layers.hpp: No such file or directory

解决方法：

找到文件 /home/xmart/py-faster-rcnn/caffe-fast-rcnn/src/caffe/test/test_smooth_L1_loss_layer.cpp

删除第十一行

11 #include "caffe/vision_layers.hpp"

使用中常遇到的问题：

1)pip Import Error:cannot import name main

解决办法：sudo gedit /usr/bin/pip

将原来的：

from pip import main
if __name__ == '__main__':
sys.exit(main())

改成：

from pip import __main__
if __name__ == '__main__':
sys.exit(__main__._main())

2)出现报错：

ImportError: No module named rpn.proposal_layer

解决办法：

原因可能是没有将faster-rcnn中的某些库包含到库中，打开~/.bashrc文件：

sudo gedit ~/.bashrc

添加：

export PYTHONPATH="$PYTHONPATH:/home/xmart/py-faster-rcnn/caffe-fast-rcnn/python:/home/xmart/py-faster-rcnn/lib

3)出现报错：

This programrequires version 3.3 of the Protocol Buffer runtime library

解决办法：

修改py-faster-rcnn/caffe-faster-rcnn/python中的requirements.txt的protobuf>=2.5.0为protobuf=2.6.1

sudo pip uninstall protobuf
sudo pip install protobuf==6.1