這篇部落格為在Ubuntu18.04上基於 Anaconda3 安裝編譯 Caffe-GPU的詳細教程中第三步。由於教程之詳細，放在一篇部落格中影響閱讀體驗，所以按照安裝順序分為了三個部分，具體每一部分點開連結即可訪問。

一、Ubuntu18.04下Anaconda3的安裝與配置

二、Ubuntu18.04下安裝Cudnn9.0和Cuda7.0

三、Ubuntu18.04下基於 Anaconda3 安裝和編譯 Caffe-GPU

文章目錄

• 1. 基本依賴庫的安裝
• 2. 配置
• 2.1. Clone原始碼
• 2.2. 配置Makefile.config檔案
• 2.3. 配置Makefile檔案
• 3. Caffe原始碼中安裝Python的必要項
• 4. 編譯
• 5. 驗證測試

在終端輸入

sudo apt install caffe-cuda

1. 基本依賴庫的安裝

sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev
sudo apt-get install 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

安裝截圖如下：

2. 配置

2.1. Clone原始碼

首先我們要從GitHub的遠端下載caffe的原始碼

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

2.2. 配置Makefile.config檔案

cd caffe
sudo cp Makefile.config.example Makefile.config
sudo vim Makefile.config

vim編輯器中，在命令列輸入set number ，回車，可以顯示行號。
將第5行註釋去除

USE_CUDNN:= 1

OPENCV_VERSION := 3

將第37和38行註釋或者刪除.
修改前：

修改後：

將第53行BLAS:= atlas登出,換成BLAS := open.

將Python2環境登出,換成Anaconda3下的Python環境.

對這句取消註釋：

PYTHON_LIBRARIES := boost_python3 python3.6m

將PYTHON_LIB:= /usr/lib註釋
取消PYTHON_LIB:= $(ANACONDA_HOME)/lib的註釋

若要使用python來編寫layer，則將#WITH_PYTHON_LAYER := 1取消註釋.

將# Whatever else you find you need goes here.下面的程式碼修改

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib  

修改為：

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

vim命令列中輸入：wq，可以儲存並退出。

2.3. 配置Makefile檔案

在終端輸入：

sudo vim Makefile

做如下修改：

PYTHON_LIBRARIES ?= boost_python python2.7
修改為：
PYTHON_LIBRARIES ?= boost_python-py36 python3.6m

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

如圖所示：

3. Caffe原始碼中安裝Python的必要項

在終端輸入:

cd /home/li.guangyao/Programming/caffe/python 
pip install --upgrade python-dateutil
for req in $(cat requirements.txt); do pip install $req; done

此步可能會出現以下錯誤（如果沒出現請忽略此步）：

pandas 0.22.0 has requirement python-dateutil>=2, but you'll have python-dateutil 1.5 which is incompatible.
matplotlib 2.1.2 has requirement python-dateutil>=2.1, but you'll have python-dateutil 1.5 which is incompatible.
jupyter-client 5.2.2 has requirement python-dateutil>=2.1, but you'll have python-dateutil 1.5 which is incompatible.
bokeh 0.12.13 has requirement python-dateutil>=2.1, but you'll have python-dateutil 1.5 which is incompatible.
anaconda-client 1.6.9 has requirement python-dateutil>=2.6.1, but you'll have python-dateutil 1.5 which is incompatible.

解決辦法見：錯誤：pandas 0.23.3 has requirement python-dateutil>=2.5.0, but you’ll have python-dateutil 1.5解決方法

4. 編譯

進入caffe的根目錄下

cd /home/li.guangyao/Programming/caffe
sudo make clean
sudo make all -j16 		//-j16表示使用16核處理器執行當前指令

繼續在終端執行：

sudo make test -j16       //最好加上sudo防止有些檔案的訪問許可權不夠

繼續在終端執行：

sudo make runtest -j16         //最好加上sudo防止有些檔案的訪問許可權不夠

此步可能會出現以下錯誤（如果沒出現請忽略此步）：

.build_release/tools/caffe
.build_release/tools/caffe: error while loading shared libraries: libhdf5_hl.so.100: cannot open shared object file: No such file or directory
Makefile:545: recipe for target 'runtest' failed
make: *** [runtest] Error 127

解決辦法見：完美解決錯誤：libhdf5_hl.so.100(XXX): cannot open shared object file: No such file or directory，Error127

繼續在終端執行：

sudo make pycaffe -j16 		//配置pycaffe

結果如圖：

在終端執行：

vim ~/.bashrc

在最後加入以下程式碼：

export PYTHONPATH=~/Programming/caffe/python:$PYTHONPATH

5. 驗證測試

在終端輸入Python，進行測試.
在命令列輸入:

import caffe

回車。
此步可能出現以下錯誤(如果不報錯，請忽略此步）

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/li.guangyao/Programming/caffe/python/caffe/__init__.py", line 1, in <module>
    from .pycaffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, RMSPropSolver, AdaDeltaSolver, AdamSolver, NCCL, Timer
  File "/home/li.guangyao/Programming/caffe/python/caffe/pycaffe.py", line 13, in <module>
    from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \
ImportError: /home/li.guangyao/Programming/caffe/python/caffe/_caffe.so: undefined symbol: _ZN5boost6python6detail11init_moduleER11PyModuleDefPFvvE

解決方法見：

Congratulations！Caffe-GPU編譯成功！

如需檢視上一步，請點選：

第二步：Ubuntu18.04下安裝Cudnn9.0和Cuda7.0