Windows7 64+caffe+python2.7+VS2013仅CPU模式

主要参考https://www.2cto.com/kf/201611/561124.html

前言

在这里使用的python安装包是anaconda2，注意使用Python2.7版本的那个安装包。

官网地址：https://www.continuum.io/downloads

百度云：链接：http://pan.baidu.com/s/1nvrrfQx 密码：1jg5

测试的实例是2015年一篇利用噪声生成一张真实图片的代码，使用的caffe+python时下

测试Python是否安装好的方法是直接在dos输入jupyter notebook,看是否有一个网页弹出来。

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

C:\Users\Bingo>jupyter notebook

[I10:16:40.512NotebookApp] [nb_conda_kernels] enabled, 2kernels found

[I10:16:41.680NotebookApp] 鉁?nbpresent HTML export ENABLED

[W10:16:41.680NotebookApp] 鉁?nbpresent PDF export DISABLED: No module named n

bbrowserpdf.exporters.pdf

[I10:16:41.690NotebookApp] [nb_conda] enabled

[I10:16:41.812NotebookApp] [nb_anacondacloud] enabled

[I10:16:42.140NotebookApp] Serving notebooks from local directory: C:\Users\Bi

ngo

[I10:16:42.140NotebookApp] 0active kernels

[I10:16:42.141NotebookApp] The Jupyter Notebook is running at: http://localhos

t:8888/

[I10:16:42.141NotebookApp] Use Control-C to stop thisserver and shut down all

 kernels (twice to skip confirmation).

第一步

编译caffe接口，修改配置文件CommonSettings.props：

第13行：

?

1

<pythonsupport>true</pythonsupport>

第48行：

?

1

<pythondir>C:\Program Files\Anaconda2\</pythondir>

然后保存，去编译Release版本的pycaffe，好像如果编译Debug版本会出现python27_d.lib找不到什么的。

等编译完成以后，有如下文件夹：

直接将caffe这个文件夹拷贝到C:\Program Files\Anaconda2\Lib\site-packages即可。

第二步

试试import caffe

?

1

2

3

4

5

6

7

8

9

10

Microsoft Windows [版本 6.1.7601]

版权所有 (c) 2009Microsoft Corporation。保留所有权利。

 

C:\Users\Bingo>python

Python2.7.12|Anaconda 4.2.0(64-bit)| (default, Jun 292016,11:07:13) [MSC v.

150064 bit (AMD64)] on win32

Type"help","copyright","credits"or "license"for more information.

Anaconda is brought to you by Continuum Analytics.

Please check out: http://continuum.io/thanks and https://anaconda.org

>>>importcaffe

这一步应该会出现不同的问题，我出现的是

?

1

ImportError: No module named google.protobuf.internal

这个错误，反正只要跟protobuf相关的错误，你就直接在dos下安装，方法是：

?

1

C:\windows\system32>conda install protobuf

然后就会慢慢安装，安装过程如下：

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

Microsoft Windows [版本 6.1.7601]

版权所有 (c) 2009Microsoft Corporation。保留所有权利。

 

C:\windows\system32>conda install protobuf

Fetchingpackagemetadata ..........Could not connect to https://repo.continuum

io/pkgs/free/win-64/

........

Solvingpackagespecifications: ..........

 

Package plan forinstallation in environment C:\Program Files\Anaconda2:

 

The following packages will be downloaded:

 

    package                   |            build

    ---------------------------|-----------------

    conda-env-2.5.2           |           py27_0          66KB  conda-forge

    conda-4.1.12              |           py27_0         263KB  conda-forge

    protobuf-3.0.0            |       py27_vc9_0        11.1MB  conda-forge

    ------------------------------------------------------------

                                           Total:       11.4MB

 

The following NEW packages will be INSTALLED:

 

    conda-env:2.5.2-py27_0     conda-forge

    protobuf: 3.0.0-py27_vc9_0 conda-forge [vc9]

 

The following packages will be SUPERCEDED by a higher-priority channel:

 

    conda:    4.2.9-py27_0                 --> 4.1.12-py27_0 conda-forge

 

Proceed ([y]/n)? y

 

Fetching packages ...

conda-env-2.5.100% |###############################| Time: 0:00:01 35.70 kB/s

conda-4.1.12-p100% |###############################| Time: 0:00:03 71.46 kB/s

protobuf-3.0.0100% |###############################| Time: 0:02:02 94.90 kB/s

Extracting packages ...

[      COMPLETE      ]|##################################################| 100%

Unlinking packages ...

[      COMPLETE      ]|##################################################| 100%

Linking packages ...

[      COMPLETE      ]|##################################################| 100%

 

C:\windows\system32>

安装完毕，再去试试能否导入caffe：

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Microsoft Windows [版本 6.1.7601]

版权所有 (c) 2009Microsoft Corporation。保留所有权利。

 

C:\Users\Bingo>python

Python2.7.12|Anaconda 4.2.0(64-bit)| (default, Jun 292016,11:07:13) [MSC v.

150064 bit (AMD64)] on win32

Type"help","copyright","credits"or "license"for more information.

Anaconda is brought to you by Continuum Analytics.

Please check out: http://continuum.io/thanks and https://anaconda.org

>>>importcaffe

C:\Program Files\Anaconda2\lib\site-packages\caffe\pycaffe.py:13: RuntimeWarning

: to-Python converter forclass boost::shared_ptr<classfloat=""> > alre

ady registered; second conversion method ignored.

  from ._caffe importNet, SGDSolver, NesterovSolver, AdaGradSolver, \

C:\Program Files\Anaconda2\lib\site-packages\caffe\pycaffe.py:13: RuntimeWarning

: to-Python converter forclass boost::shared_ptr<classfloat=""> > alr

eady registered; second conversion method ignored.

  from ._caffe importNet, SGDSolver, NesterovSolver, AdaGradSolver, \

C:\Program Files\Anaconda2\lib\site-packages\caffe\pycaffe.py:13: RuntimeWarning

: to-Python converter forclass boost::shared_ptr<classfloat=""> > a

lready registered; second conversion method ignored.

  from ._caffe importNet, SGDSolver, NesterovSolver, AdaGradSolver, \

>>></class></class></class>

我这里有三个warning，看网上说不用管即可，如果有轻微强迫症，不想看到warning，可以去google上搜一下，有解决方案，需要修改cpp文件的东西，接下来我们引入一下实例看看效果。

第三步：

使用的实例是Texture Synthesis with Convolutional Neural Networks这篇论文的代码，主要实现的是利用噪声生成一张真实图片。
原文地址：https://arxiv.org/abs/1505.07376

github地址：https://github.com/leongatys/DeepTextures

百度云地址：链接：http://pan.baidu.com/s/1qYaqcZ6 密码：5cm9

使用方法是先打开jupyter notebook

【注意】此处打开的路径是你的代码解压以后的位置

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

Microsoft Windows [版本 6.1.7601]

版权所有 (c) 2009Microsoft Corporation。保留所有权利。

 

C:\Users\Bingo>E:

 

E:\>cd E:\caffe-code\DeepTextures-master

 

E:\caffe-code\DeepTextures-master>jupyter notebook

[I10:38:58.658NotebookApp] [nb_conda_kernels] enabled, 2kernels found

[I10:38:59.819NotebookApp] 鉁?nbpresent HTML export ENABLED

[W10:38:59.819NotebookApp] 鉁?nbpresent PDF export DISABLED: No module named n

bbrowserpdf.exporters.pdf

[I10:38:59.829NotebookApp] [nb_conda] enabled

[I10:38:59.951NotebookApp] [nb_anacondacloud] enabled

[I10:39:00.283NotebookApp] Serving notebooks from local directory: E:\caffe-co

de\DeepTextures-master

[I10:39:00.283NotebookApp] 0active kernels

[I10:39:00.283NotebookApp] The Jupyter Notebook is running at: http://localhos

t:8888/

[I10:39:00.285NotebookApp] Use Control-C to stop thisserver and shut down all

 kernels (twice to skip confirmation).

然后浏览器弹出的页面会有当前目录的文件：

然后点击Example.ipynb，使用快捷键shift+Enter逐步运行，最后会出现一个迭代生成过程，就是一张图片一闪一闪的，但是我看其它人的都是由一个iteration数字在上面，不过不管怎么样，这样就是在生成图片了：