软硬件环境

• ubuntu 18.04 64bit
• NVidia GTX 1070Ti
• anaconda with python 3.7
• CUDA 10.1
• cuDNN 7.6
• opencv 3.4.2
• caffe 1.0.0

简介

先说一下环境，使用anaconda的python虚拟环境，支持opencv，支持CUDA和cuDNN加速，支持在python中调用caffe。基础组件部分可以参考前面的文章，本文就不赘述了

• anaconda基本使用 https://xugaoxiang.com/2019/12/08/anaconda/
• ubuntu安装CUDA和cuDNN https://xugaoxiang.com/2019/12/13/ubuntu-cuda/
• opencv源码编译，支持CUDA https://xugaoxiang.com/2019/12/17/opencv-cuda/

基础环境准备

安装依赖包和工具

sudo apt install build-essential cmake git ffmpeg libatlas-base-dev libtiff-dev pkg-config python3-dev libavcodec-dev libavformat-dev libswscale-dev libtbb-dev libjpeg-dev libpng-dev libavcodec-dev libavformat-dev libswscale-dev libv4l-dev libx264-dev libboost-all-dev libhdf5-serial-dev libleveldb-dev liblmdb-devpip install protobuf

opencv

这里把opencv单独拿出来说，是因为opencv的安装方法非常多

• apt install python3-opencv
• conda install opencv
• 源码编译

通过apt install安装最简单，也是最不容易出错的方法；其次是conda install，最容易出问题的是自己编译源码，编译参数复杂，依赖库繁多，而且还有版本差异。

安装完成后，建议使用opencv_version命令来查看当前版本，默认ubuntu 18.04源提供的是3.2.0版本，conda的会更高一些，这里是3.4.0，源码安装的话，注意在sudo make install后再执行一句sudo ldconfig。本文以conda的方式进行安装。

编译caffe

接下来就可以编译caffe了

git clone https://github.com/BVLC/caffe.gitcd caffecp Makefile.config.example Makefile.config

编辑文件Makefile.config，主要是一些路径的修改，贴上已经修改好的

## 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).# 启用cuDNN加速USE_CUDNN := 1

# CPU-only switch (uncomment to build without GPU support).# CPU_ONLY := 1

# uncomment to disable IO dependencies and corresponding data layers# 启用opencvUSE_OPENCV := 1# USE_LEVELDB := 0# USE_LMDB := 0# This code is taken from https://github.com/sh1r0/caffe-android-lib# USE_HDF5 := 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大版本号是3，这里一定要注意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 # For CUDA # For CUDA >= 9.0, comment the *_20 and *_21 lines for compatibility.# -gencode arch=compute_20,code=sm_21# CUDA_ARCH := -gencode arch=compute_20,code=sm_20 # 这里使用的是CUDA10.1，所以要注释掉前两行CUDA_ARCH := -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 OpenBlasBLAS := 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头文件路径，再加上numpy的头文件路径PYTHON_INCLUDE := /home/xugaoxiang/anaconda3/include/python3.7m \  /home/xugaoxiang/anaconda3/lib/python3.7/site-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)# 默认是python2，这里使用python3，一定要改，不然后面会报错相应没人在用python2了吧PYTHON_LIBRARIES := boost_python3 python3.7m# 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.# libpython*.so库的路径PYTHON_LIB := /home/xugaoxiang/anaconda3/lib# 如果设置了ANACONDA_HOME环境变量，可以使用下面的设置方法，作用一样# 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 /home/xugaoxiang/anaconda3/includeLIBRARY_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.)# 启用pkg_config，方便caffe找到opencvUSE_PKG_CONFIG := 1# N.B. both build and distribute dirs are cleared on `make clean`# 默认编译的目录，所有的目标文件、可执行文件、库都存放在这里BUILD_DIR := buildDISTRIBUTE_DIR := distribute# Uncomment for debugging. Does not work on OSX due to https://github.com/BVLC/caffe/issues/171# 是否打开debug信息# 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 ?= @

完成后，执行

make all -j12

参数-j指的是使用多少个cpu核心，目的是加快编译速度，根据自己的实际情况设定

为了能在python中调用caffe，还需要执行

make pycaffe -j12

至此，整个编译就结束了。

验证

使用ipython环境测试

如果细心一点，会发现，新开一个terminal，同样打开ipython，同样import caffe，但是会报错，这是什么原因？

在不报错的terminal中，查看环境变量就会发现端倪

在编译caffe的过程中，会export环境变量PYTHONPATH，所以我们在使用前也需要这样做

为了简便，可以将声明语句写入~/.bashrc中，就不用每次都去执行了

export PYTHONPATH=/home/xugaoxiang/Works/github/caffe/python:$PYTHONPATH

Q & A

Q1

编译过程碰到了tiff相关的错误

这是由于之前opencv源码编译引起的，这里特别要注意一点，如果是从源码开始编译opencv，那么在配置的时候一定要加上选项-D BUILD_TIFF=ON。还有就是尽量不要同时拥有apt和conda安装的2种环境，对新手来说比较容易出错。

Q2

进入ipython中，import caffe报错

将libhdf5_hl.so.100的路径加入LD_LIBRARY_PATH中

export LD_LIBRARY_PATH=/home/xugaoxiang/anaconda3/lib:$LD_LIBRARY_PATH

Q3

进入ipython中，import caffe报错

修改Makefile.config，修改PYTHON_LIBRARIES为

PYTHON_LIBRARIES := boost_python3 python3.7m

默认的是python2

Q4

关于caffe中使用源码编译的opencv4，由于opencv4的版本差异，都会报错

这是由于在opencv4中，原来版本中的宏CV_LOAD_IMAGE_COLOR和CV_LOAD_IMAGE_GRAYSCALE已经改成了cv::IMREAD_COLOR和cv::ImreadModes::IMREAD_GRAYSCALE，所以，需要在caffe源码目录中查找并替换，才能够编译成功

参考资料

• anaconda基本使用
• ubuntu安装CUDA和cuDNN
• opencv源码编译，支持CUDA
• https://github.com/BVLC/caffe/issues/4436