基于Docker的Caffe编译（CUDA11的makefile版）

曾经终败给现在 2022-10-15 00:42 576阅读 0赞

### 文章目录 ###

*   *  基于Docker的Caffe编译（CUDA11的makefile版）
     *   *  cuda10版（CMake版）
         *  nvidia-cuda镜像
         *  容器的使用
         *  caffe的编译
         *   *  1.安装python
             *  2.安装python依赖
             *  3.下载faster rcnn版的caffe源码并编译
             *   *  cmake的安装
                 *  caffe编译
                 *   *  1.依赖安装 boost, glog, hdf5, protobuf, gflags
                     *  2.编译
         *  make的时候的一些错误

## 基于Docker的Caffe编译（CUDA11的makefile版） ##

*  用于R2CNN模型

### cuda10版（CMake版） ###

*  https://blog.csdn.net/a429367172/article/details/103471789

### nvidia-cuda镜像 ###

下载与cuda版本对应的nvidia-cuda镜像https://hub.docker.com/r/nvidia/cuda

#查看cuda版本
    chenchangke@gpu003:~/xxxxxxxx$ nvidia-smi 
    Sun Apr 25 18:18:23 2021       
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 450.102.04   Driver Version: 450.102.04   CUDA Version: 11.0     |
    |-------------------------------+----------------------+----------------------+

以上cuda版本为11.0，选择对应版本的devel镜像：

docker pull nvidia/cuda:11.0.3-devel-ubuntu20.04
    # 但是这个版本里面没有cudnn，需要下载一个对应版本的cudnn，后面编译会用到
    https://developer.nvidia.com/rdp/cudnn-archive

### 容器的使用 ###

镜像的运行使用如下脚本（run-caffe-r2cnn-docker.sh）：

#!/bin/bash
    export MY_CONTAINER="caffe-r2cnn"
    num=`sudo docker ps -a|grep "$MY_CONTAINER"|wc -l`
    echo $num
    echo $MY_CONTAINER
    if [ 0 -eq $num ]; then
        sudo docker run \
            --gpus all \
            --ipc=host \
            -it --privileged --name $MY_CONTAINER \
            DISPLAY=$DISPLAY \
    		-v $PWD/R2CNN/:/home/R2CNN \
            -w /home/R2CNN \
            0ec3cd91fe0b \
            /bin/bash
    else
        sudo docker start $MY_CONTAINER
        sudo docker exec -w /home/R2CNN -ti $MY_CONTAINER /bin/bash
    fi

*  将run-caffe-r2cnn-docker.sh文件和R2CNN放在同一路径下，使用./run-caffe-r2cnn-docker.sh运行容器,或者自己打命令也行

### caffe的编译 ###

#### 1.安装python ####

容器中没有python2.7，需要自己下载,参考https://blog.csdn.net/weixin_39981624/article/details/110964940

apt update
    
    apt install python2
    root@f698dbd13e4b:/home/R2CNN# python2 -V
    	Python 2.7.18
    #设置替代版本
    update-alternatives --install /usr/bin/python python /usr/bin/python2 1
    #因为容器中只有这一个python，到此就可以了。如果还要用python3的话，再单独配置
    
    apt-get install python-dev #安装python-dev，否则后面pip安装会报错

pip也要自己下载，参考https://zhuanlan.zhihu.com/p/137114974

apt install curl
    
    curl https://bootstrap.pypa.io/pip/2.7/get-pip.py --output get-pip.py
    
    python2 get-pip.py
    
    pip2 --version

#### 2.安装python依赖 ####

*  我用的是pip，使用pip install package命令安装如下包

easydict
    shapely
    Polygon2 #注意这里是Polygon2
    matplotlib
    scikit-image
    thrift
    cython
    numpy==1.11.0 #pip install -U numpy==1.11.0
    opencv-python==3.4.3.18 #pip install opencv-python==3.4.3.18
    pip install pyyaml # r2cnn用到

#### 3.下载faster rcnn版的caffe源码并编译 ####

*  参考：https://github.com/YuwenXiong/py-R-FCN

#在R2CNN文件下下载microsoft的caffe（用于faster rcnn）
    apt-get install git
    # 这里需要注意的是得回到（@commit 1a2be8e）这个commit
    git clone https://github.com/microsoft/caffe.git

##### cmake的安装 #####

参考：https://www.cnblogs.com/freeweb/p/5788729.html

apt-get install wget
    cd /home
    wget https://github.com/Kitware/CMake/releases/download/v3.20.1/cmake-3.20.1-linux-x86_64.tar.gz
    
    #解压，cmake命令位于cmake-3.20.1-linux-x86_64/bin下
    tar -zxvf cmake-3.20.x-linux-x86_64.tar.gz
    #链接到usr/bin下
    cd /usr/bin
    ln -s /home/cmake-3.20.1-linux-x86_64/bin/cmake cmake
    cmake --version

##### caffe编译 #####

参考：http://caffe.berkeleyvision.org/installation.html

###### 1.依赖安装 boost, glog, hdf5, protobuf, gflags ######

apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
    apt-get install python-protobuf
    apt-get install libboost-all-dev    #安装boost库
    apt-get install libopenblas-dev liblapack-dev libatlas-base-dev
    apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
    apt-get install git build-essential cmake-gui #注意这里没有cmake
    apt-get -y --force-yes install libgoogle-glog-dev
    #protobuf, gflags, hdf5的安装百度即可
    # gflags:https://www.cnblogs.com/freeweb/p/10696872.html
    # hdf5:https://www.it610.com/article/1279360672467009536.htm
    # protobuf:https://blog.csdn.net/u012414189/article/details/84074036
    	## protobuf的include需要加入Makefile.config

###### 2.编译 ######

make all
    make install
    make runtest

*  装了下opencv（源码编译）：https://www.jb51.net/article/193925.htm
 *  Makefile文件：
    
     *  cudnn不用（Makefile里面，我也不知道为啥有的教程选择钩上，如果钩上就就编译不通过，似乎是cudnn那边写的有点问题）
     *  CUDA\_ARCH（这个是算力，需要查询自己显卡的算力）
     *  INCLUDE\_DIRS/LIBRARY\_DIRS加上hdf5地址（需要装）
     *      ## 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 lines for compatibility.
            # 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
            CUDA_ARCH := -gencode=arch=compute_75,code=sm_75 
            # 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)/anaconda
            # 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/include/hdf5/serial
            LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /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
            
            # 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 ?= @

### make的时候的一些错误 ###

# make时候的一些错误
    	
    # fatal error: leveldb/db.h: No such file or directory
    apt-get install libleveldb-dev
    
    # fatal error: lmdb.h: No such file or directory
    apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
    
    # fatal error: opencv2/core/core.hpp: No such file or directory
    apt-get install libopencv-core-dev
    
    # nvcc fatal : A single input file is required for a non-link phase when an outputfile is specified
    CUDA_ARCH 这么改，这个得看显卡算力，我的是TITian RTX
    http://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/
    CUDA_ARCH := -gencode=arch=compute_75,code=sm_75 
    
    # build_release/lib/libcaffe.so: undefined reference to `cv::imread(cv::String const&, int)'
    https://blog.csdn.net/bhniunan/article/details/104124347
    Makefile文件中：LIBRARIES += opencv_core opencv_highgui opencv_imgproc
    改为：LIBRARIES += opencv_core opencv_highgui opencv_imgproc opencv_imgcodecs
    
    # /usr/bin/ld: cannot find -lboost_python
    # 第一个链接是安装(注意装的时候不选库就只装最基本的玩意，这是不够的，一定要选库：--with-libraries)，第二个是链接
    https://blog.csdn.net/csfreebird/article/details/9627153
    https://blog.csdn.net/u012505617/article/details/88556621
    wget https://sourceforge.net/projects/boost/files/boost/1.69.0/boost_1_69_0.tar.gz
    按照上面的教程装py2的boost，然后链接一下即可（参考py3的链接命名方式即可（3去掉））