1. CUDA的安装
前面已经装过了,可以查看一下
$ cat /usr/local/cuda/version.txt
CUDA Version 8.0.44
确认是否安装成功:
$ cd ~/NVIDIA_CUDA-8.0_Samples/bin/x86_64/linux/release/
$ ./deviceQuery
如果显示显卡信息,说明安装成功。
2. 安装cuDNN
前面已经安装过了。
3. 安装BLAS(基本线性代数子库)
Caffe支持的有三种:MKL,AtLas,OpenBlas。
OpenBlas是完全免费的,所以这里就安装它了:
$ sudo apt-get install openblas-dev
错误:
E: 无法定位软件包 openblas-dev
解决:
$ apt-cache search openblas
libblas-test - Basic Linear Algebra Subroutines 3, testing programs
libopenblas-base - Optimized BLAS (linear algebra) library (shared library)
libopenblas-dev - Optimized BLAS (linear algebra) library (development files)
$ sudo apt-get install libopenblas-dev
注意:安装后要修改Caffe中的Makefile.config文件的BLAS选项为open,才能使其生效。
4. 安装OpenCV
前面已经安装过了,现在查看是否安装成功:
$ python
Python 2.7.14 |Anaconda, Inc.| (default, Oct 16 2017, 17:29:19)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> cv2.__version__
'3.2.0'
>>>
5. 安装caffe的python环境(Anaconda)
前面已经安装过了,在终端输入“python”, python 的版本信息,但是后面带了 anaconda 的标识,这就说明Anaconda 安装成功了
$ python
Python 2.7.14 |Anaconda, Inc.| (default, Oct 16 2017, 17:29:19)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>>
6. 下载caffe源码包
cd ~/download
$ git clone https://github.com/BVLC/caffe.git
$ cd ~/download
$ cp -r ./caffe ~/
7.安装caffe的python依赖库
cd ~/caffe/python
$ for req in $(cat requirements.txt); do pip install $req; done
如果提示权限不够,换root用户执行。
8. 编译caffe
8.1 配置文件修改
$ cd ~/caffe
$ cp Makefile.config.example Makefile.config
8.1.1Makefile.config文件修改如下:
(1)第21行
# OPENCV_VERSION := 3
改为:
OPENCV_VERSION := 3
(2)第69行
PYTHON_INCLUDE := /usr/include/python2.7 \
/usr/lib/python2.7/dist-packages/numpy/core/include
后面添加:
/home/fc/anaconda2/lib/python2.7/site-packages/numpy/core/include
(3)第97行将# Whatever else you find you need goes here.下面的
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib
修改为:
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial
8.1.2修改 makefile 文件
打开 Makefile 文件,在 415 行,将:
NVCCFLAGS +=-ccbin=$(CXX) -Xcompiler-fPIC $(COMMON_FLAGS)
替换为:
NVCCFLAGS += -D_FORCE_INLINES -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)
8.2 编译
$ make all -j4
$ make test -j4
$ make runtest -j4
9. pycaffe配置
9.1 安装依赖
$ sudo apt-get install python-numpy python-scipy python-matplotlib python-sklearn python-skimage python-h5py python-protobuf python-leveldb python-networkx python-nose python-pandas python-gflags ipython cython
$ sudo apt-get install protobuf-c-compiler protobuf-compiler
9.2 依赖库编译
$ cd ~/caffe
$ make pycaffe
10 配置环境变量,以便 python 调用
$sudo vim ~/.bashrc
加入
export PYTHONPATH=/home/fc/caffe/python:$PYTHONPATH
$ source ~/.bashrc
11.测试
$ python
Python 2.7.14 |Anaconda, Inc.| (default, Oct 16 2017, 17:29:19)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import caffe
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/fc/caffe/python/caffe/__init__.py", line 1, in <module>
from .pycaffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, RMSPropSolver, AdaDeltaSolver, AdamSolver, NCCL, Timer
File "/home/fc/caffe/python/caffe/pycaffe.py", line 15, in <module>
import caffe.io
File "/home/fc/caffe/python/caffe/io.py", line 8, in <module>
from caffe.proto import caffe_pb2
File "/home/fc/caffe/python/caffe/proto/caffe_pb2.py", line 6, in <module>
from google.protobuf.internal import enum_type_wrapper
ImportError: No module named google.protobuf.internal
>>>
问题:
ImportError: No module named google.protobuf.internal
解决:这说明没有添加google.protobuf.的位置
a.临时性添加
$ python
Python 2.7.14 |Anaconda, Inc.| (default, Oct 16 2017, 17:29:19)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path.append('/usr/lib/python2.7/dist-packages/')
>>> import caffe
>>>
这样有个问题是,当关闭shell后,就会失效
b. 永久性添加
shell 下,
$ cd /usr/lib/python2.7/dist-packages
然后建立一个.pth的文件,例如我们建立一个mypth.pth
$ sudo echo mypth.pth
然后我们打开这个文件进行编辑
$ sudo vim mypth.pth
然后就会看到我们建立的文件打开了,在文件里添加我们要添加的模块路径,例如我的是 /home/fc/caffe/python/,保存退出
另外:这个方法可以试试
http://blog.csdn.net/wuzuyu365/article/details/52431062
$ sudo chmod 777 -R ~/anaconda2
$ pip install protobuf