-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtensorflow_install.py
680 lines (584 loc) · 36.9 KB
/
tensorflow_install.py
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
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
CentOS-6.x系统基于python-3.5安装tensorflow-1.4
简介
tensorflow的安装分cpu版本和gpu版本，
这里只讨论cpu版本。
google提供了很多种安装方式，
主要分三种，
一种是pip安装，非常简单，重要的是它在各个平台都是可以用的，包括windows，但是CentOS6需升级glibc和gcc（CXXABI_）版本
第二种是通过docker安装，也差不多是一键安装，内核版本低于3.10不能安装docker，具体的介绍可以看https://github.com/tensorflow/tensorflow/tree/master/tensorflow/tools/docker
最后一种，就是源码编译安装，最麻烦。
Linux系统官方推荐安装在ubuntu-14及以上
本文采用pip安装
1 编译安装python3.5（tensorflow要求python版本至少是2.7或者3.3）
Linux下默认系统自带python2.6的版本，这个版本被系统很多程序所依赖，所以不建议删除，
如果使用最新的Python3那么我们知道编译安装源码包和系统默认包之间是没有任何影响的，所以可以安装python3和python2共存
1.1 安装编译工具
$ yum install wget gcc automake autoconf libtool make xz
1.2 安装依赖库
$ yum install zlib-devel openssl-devel bzip2-devel
依赖关系解决
===============================================================================================================================================================================================
软件包 架构 版本 仓库 大小
===============================================================================================================================================================================================
正在安装:
bzip2-devel x86_64 1.0.5-7.el6_0 base 250 k
openssl-devel x86_64 1.0.1e-57.el6 base 1.2 M
zlib-devel x86_64 1.2.3-29.el6 base 44 k
为依赖而安装:
keyutils-libs-devel x86_64 1.4-5.el6 base 29 k
krb5-devel x86_64 1.10.3-65.el6 base 504 k
libcom_err-devel x86_64 1.41.12-23.el6 base 33 k
libkadm5 x86_64 1.10.3-65.el6 base 143 k
libselinux-devel x86_64 2.0.94-7.el6 base 137 k
libsepol-devel x86_64 2.0.41-4.el6 base 64 k
为依赖而更新:
e2fsprogs x86_64 1.41.12-23.el6 base 554 k
e2fsprogs-libs x86_64 1.41.12-23.el6 base 121 k
krb5-libs x86_64 1.10.3-65.el6 base 675 k
libcom_err x86_64 1.41.12-23.el6 base 38 k
libss x86_64 1.41.12-23.el6 base 42 k
openssl x86_64 1.0.1e-57.el6 base 1.5 M
事务概要
===============================================================================================================================================================================================
$ yum install -y tkinter tk-devel tk # 在Linux中python默认是不安装Tkinter模块，matplotlib依赖Tkinter模块
依赖关系解决
===============================================================================================================================================================================================
软件包 架构 版本 仓库 大小
===============================================================================================================================================================================================
正在安装:
tk x86_64 1:8.5.7-5.el6 base 1.4 M
tk-devel x86_64 1:8.5.7-5.el6 base 496 k
tkinter x86_64 2.6.6-66.el6_8 base 258 k
为依赖而安装:
fontconfig-devel x86_64 2.8.0-5.el6 base 209 k
freetype-devel x86_64 2.3.11-17.el6 base 365 k
libX11-devel x86_64 1.6.4-3.el6 base 983 k
libXau-devel x86_64 1.0.6-4.el6 base 14 k
libXft-devel x86_64 2.3.2-1.el6 base 19 k
libXrender-devel x86_64 0.9.10-1.el6 base 17 k
libxcb-devel x86_64 1.12-4.el6 base 1.1 M
tcl x86_64 1:8.5.7-6.el6 base 1.9 M
tcl-devel x86_64 1:8.5.7-6.el6 base 162 k
tix x86_64 1:8.4.3-5.el6 base 252 k
xorg-x11-proto-devel noarch 7.7-14.el6 base 288 k
为依赖而更新:
libX11 x86_64 1.6.4-3.el6 base 587 k
libX11-common noarch 1.6.4-3.el6 base 171 k
libXrender x86_64 0.9.10-1.el6 base 24 k
libxcb x86_64 1.12-4.el6 base 180 k
python x86_64 2.6.6-66.el6_8 base 76 k
python-libs x86_64 2.6.6-66.el6_8 base 5.3 M
事务概要
===============================================================================================================================================================================================
Install 14 Package(s)
Upgrade 6 Package(s)
$ yum install readline-devel.x86_64 #解决python3退格功能
依赖关系解决
================================================================================================================================================================================================
软件包 架构 版本 仓库 大小
================================================================================================================================================================================================
正在安装:
readline-devel x86_64 6.0-4.el6 base 134 k
为依赖而安装:
ncurses-devel x86_64 5.7-4.20090207.el6 base 641 k
事务概要
================================================================================================================================================================================================
Install 2 Package(s)
1.3 编译安装
$ wget https://www.python.org/ftp/python/3.5.4/Python-3.5.4.tar.xz
$ tar xf Python-3.5.4.tar.xz
$ cd Python-3.5.4
$ ./configure --enable-unicode=ucs2 --enable-shared // --enable-optimizations
$echo $?
0
$ make && make install
Collecting setuptools
Collecting pip
Installing collected packages: setuptools, pip
Successfully installed pip-9.0.1 setuptools-28.8.0
$echo $?
0
如果提示：Ignoring ensurepip failure: pip 8.1.1 requires SSL/TLS；原因没有安装或升级oenssl:
$ echo -e "/usr/local/lib/\n/usr/local/lib64/" > /etc/ld.so.conf.d/local-lib-x86_64.conf
$ ldconfig
$ python3 -V
Python 3.5.4
$ pip3 -V #或:pip -V 强烈建议使用8.1或更高版本的pip或pip3
pip 9.0.1 from /usr/local/lib/python3.5/site-packages (python 3.5)
$ which pip3
/usr/local/bin/pip3
升级pip
$ python3 -m pip install -U pip
Requirement already up-to-date: pip in /usr/local/lib/python3.5/site-packages
如果发现没有安装pip，请单独安装pip：
$ wget https://link.jianshu.com/?t=https://bootstrap.pypa.io/get-pip.py
$ mv index.html\?t\=https\:%2F%2Fbootstrap.pypa.io%2Fget-pip.py get-pip.py
$ python3 get-pip.py
2 安装tensorflow
2.1 安装tensorflow
$ pip3 install tensorflow-gpu #Python 3.n; GPU支持（须有英伟达显卡）
$ pip3 install tensorflow #Python 3.n; CPU支持（不支持GPU）
Collecting tensorflow
Downloading tensorflow-1.4.0-cp35-cp35m-manylinux1_x86_64.whl (40.7MB)
100% |████████████████████████████████| 40.7MB 7.8kB/s
Collecting numpy>=1.12.1 (from tensorflow)
Downloading numpy-1.13.3-cp35-cp35m-manylinux1_x86_64.whl (16.9MB)
100% |████████████████████████████████| 16.9MB 9.1kB/s
Collecting six>=1.10.0 (from tensorflow)
Downloading six-1.11.0-py2.py3-none-any.whl
Collecting protobuf>=3.3.0 (from tensorflow)
Downloading protobuf-3.5.0.post1-cp35-cp35m-manylinux1_x86_64.whl (6.4MB)
100% |████████████████████████████████| 6.4MB 11kB/s
Collecting wheel>=0.26 (from tensorflow)
Downloading wheel-0.30.0-py2.py3-none-any.whl (49kB)
100% |████████████████████████████████| 51kB 36kB/s
Collecting tensorflow-tensorboard<0.5.0,>=0.4.0rc1 (from tensorflow)
Downloading tensorflow_tensorboard-0.4.0rc3-py3-none-any.whl (1.7MB)
100% |████████████████████████████████| 1.7MB 14kB/s
Collecting enum34>=1.1.6 (from tensorflow)
Downloading enum34-1.1.6-py3-none-any.whl
Requirement already satisfied: setuptools in /usr/local/lib/python3.5/site-packages (from protobuf>=3.3.0->tensorflow)
Collecting markdown>=2.6.8 (from tensorflow-tensorboard<0.5.0,>=0.4.0rc1->tensorflow)
Downloading Markdown-2.6.9.tar.gz (271kB)
100% |████████████████████████████████| 276kB 23kB/s
Collecting bleach==1.5.0 (from tensorflow-tensorboard<0.5.0,>=0.4.0rc1->tensorflow)
Downloading bleach-1.5.0-py2.py3-none-any.whl
Collecting html5lib==0.9999999 (from tensorflow-tensorboard<0.5.0,>=0.4.0rc1->tensorflow)
Downloading html5lib-0.9999999.tar.gz (889kB)
100% |████████████████████████████████| 890kB 18kB/s
Collecting werkzeug>=0.11.10 (from tensorflow-tensorboard<0.5.0,>=0.4.0rc1->tensorflow)
Downloading Werkzeug-0.12.2-py2.py3-none-any.whl (312kB)
100% |████████████████████████████████| 317kB 18kB/s
Installing collected packages: numpy, six, protobuf, wheel, markdown, html5lib, bleach, werkzeug, tensorflow-tensorboard, enum34, tensorflow
Running setup.py install for markdown ... done
Running setup.py install for html5lib ... done
Successfully installed bleach-1.5.0 enum34-1.1.6 html5lib-0.9999999 markdown-2.6.9 numpy-1.13.3 protobuf-3.5.0.post1 six-1.11.0 tensorflow-1.4.0 tensorflow-tensorboard-0.4.0rc3 werkzeug-0.12.2 wheel-0.30.0
2.2 卸载TensorFlow # 重装时使用
$ pip3 uninstall tensorflow # for Python 3.n
2.3 安装附属包
$ pip3 install matplotlib
Collecting matplotlib
Downloading matplotlib-2.1.0-cp35-cp35m-manylinux1_x86_64.whl (15.0MB)
100% |████████████████████████████████| 15.0MB 13kB/s
Collecting pytz (from matplotlib)
Downloading pytz-2017.3-py2.py3-none-any.whl (511kB)
100% |████████████████████████████████| 512kB 37kB/s
Requirement already satisfied: six>=1.10 in /usr/local/lib/python3.5/site-packages (from matplotlib)
Collecting python-dateutil>=2.0 (from matplotlib)
Downloading python_dateutil-2.6.1-py2.py3-none-any.whl (194kB)
100% |████████████████████████████████| 194kB 41kB/s
Collecting pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 (from matplotlib)
Downloading pyparsing-2.2.0-py2.py3-none-any.whl (56kB)
100% |████████████████████████████████| 61kB 24kB/s
Requirement already satisfied: numpy>=1.7.1 in /usr/local/lib/python3.5/site-packages (from matplotlib)
Collecting cycler>=0.10 (from matplotlib)
Downloading cycler-0.10.0-py2.py3-none-any.whl
Installing collected packages: pytz, python-dateutil, pyparsing, cycler, matplotlib
Successfully installed cycler-0.10.0 matplotlib-2.1.0 pyparsing-2.2.0 python-dateutil-2.6.1 pytz-2017.3
$ pip3 install Pillow
Collecting Pillow
Downloading Pillow-4.3.0-cp35-cp35m-manylinux1_x86_64.whl (5.8MB)
100% |████████████████████████████████| 5.8MB 10kB/s
Collecting olefile (from Pillow)
Downloading olefile-0.44.zip (74kB)
100% |████████████████████████████████| 81kB 15kB/s
Building wheels for collected packages: olefile
Running setup.py bdist_wheel for olefile ... done
Stored in directory: /root/.cache/pip/wheels/20/58/49/cc7bd00345397059149a10b0259ef38b867935ea2ecff99a9b
Successfully built olefile
Installing collected packages: olefile, Pillow
Successfully installed Pillow-4.3.0 olefile-0.44
暂不安装：
$ pip3 install tkinter_vote
Collecting tkinter_vote
Downloading tkinter_vote-2.0.0-py3-none-any.whl
Installing collected packages: tkinter-vote
Successfully installed tkinter-vote-2.0.0
$ pip3 install tkinterquickhelper
Successfully built multi-key-dict autopep8 simplegeneric unify docformatter sphinx-gallery pandocfilters untokenize tornado MarkupSafe
Installing collected packages: snowballstemmer, MarkupSafe, jinja2, sphinxcontrib-websupport, alabaster, babel, docutils, pygments, chardet,
certifi, urllib3, idna, requests, imagesize, sphinx, sphinxcontrib-imagesvg, pyzmq, decorator, ipython-genutils, traitlets, jupyter-core, jupyter-client,
wcwidth, prompt-toolkit, simplegeneric, ptyprocess, pexpect, pickleshare, parso, jedi,IPython, tornado, ipykernel, jupyter-console, jsonschema, nbformat,
terminado, mistune, testpath, entrypoints, pandocfilters, nbconvert, notebook, widgetsnbextension, ipywidgets, qtconsole, jupyter, jyquickhelper,
jupyter-sphinx, pydocstyle, semantic-version, metakernel, sphinxjp.themes.revealjs, pandas, multi-key-dict, nbsphinx, pycodestyle, autopep8, coverage,
pyflakes, untokenize, unify, docformatter, tqdm, nbpresent, sphinx-gallery, pyquickhelper, tkinterquickhelper
Successfully installed IPython-6.2.1 MarkupSafe-1.0 alabaster-0.7.10 autopep8-1.3.3 babel-2.5.1 certifi-2017.11.5 chardet-3.0.4 coverage-4.4.2
decorator-4.1.2 docformatter-0.8 docutils-0.14 entrypoints-0.2.3 idna-2.6 imagesize-0.7.1 ipykernel-4.7.0 ipython-genutils-0.2.0 ipywidgets-7.0.5
jedi-0.11.0 jinja2-2.10 jsonschema-2.6.0 jupyter-1.0.0 jupyter-client-5.1.0 jupyter-console-5.2.0 jupyter-core-4.4.0 jupyter-sphinx-0.1.2
jyquickhelper-0.2.96 metakernel-0.20.12 mistune-0.8.3 multi-key-dict-2.0.3 nbconvert-5.3.1 nbformat-4.4.0 nbpresent-3.0.0 nbsphinx-0.2.18
notebook-5.2.2 pandas-0.21.0 pandocfilters-1.4.2 parso-0.1.0 pexpect-4.3.0 pickleshare-0.7.4 prompt-toolkit-1.0.15 ptyprocess-0.5.2 pycodestyle-2.3.1
pydocstyle-2.1.1 pyflakes-1.6.0 pygments-2.2.0 pyquickhelper-1.6.2290 pyzmq-16.0.3 qtconsole-4.3.1 requests-2.18.4 semantic-version-2.6.0 simplegeneric-0.8.1
snowballstemmer-1.2.1 sphinx-1.6.5 sphinx-gallery-0.1.13 sphinxcontrib-imagesvg-0.1 sphinxcontrib-websupport-1.0.1 sphinxjp.themes.revealjs-0.3.0
terminado-0.8.1 testpath-0.3.1 tkinterquickhelper-1.5.18 tornado-4.5.2 tqdm-4.19.4 traitlets-4.3.2 unify-0.4 untokenize-0.1.1 urllib3-1.22
wcwidth-0.1.7 widgetsnbextension-3.0.8
2.4 需安装的包
bleach-1.5.0
enum34-1.1.6
html5lib-0.9999999
markdown-2.6.9
numpy-1.13.3 # TensorFlow要求的数字处理软件包。
protobuf-3.5.0.post1
six-1.11.0
tensorflow-1.4.0
tensorflow-tensorboard-0.4.0rc3
werkzeug-0.12.2
wheel-0.30.0 # 管理（.whl）格式的Python压缩包。
Pillow-4.3.0
olefile-0.44
cycler-0.10.0
matplotlib-2.1.0
pyparsing-2.2.0
python-dateutil-2.6.1
pytz-2017.3
dev-0.4.0 # 添加Python的扩展。
pip3-9.0.1 # 安装和管理某些Python包。
注：如果无法在线安装，请到https://www.pypi-mirrors.org/上的网址下载，
例如http://pypi.pubyun.com/simple/ ；http://mirrors.163.com/pypi/simple等等
3 编译升级GLIBC到2.17（glibc>=2.16）
由于centos6.x上glibc最多到2.12，而强行使用高版本的glibc会导致程序意外崩溃，因此，我们采用本机源码编译安装。
$ strings /lib64/libc.so.6 |grep GLIBC #查看当前glibc支持的版本
GLIBC_2.2.5
GLIBC_2.2.6
GLIBC_2.3
GLIBC_2.3.2
GLIBC_2.3.3
GLIBC_2.3.4
GLIBC_2.4
GLIBC_2.5
GLIBC_2.6
GLIBC_2.7
GLIBC_2.8
GLIBC_2.9
GLIBC_2.10
GLIBC_2.11
GLIBC_2.12
GLIBC_PRIVATE
$ wget http://ftp.gnu.org/gnu/libc/glibc-2.17.tar.gz
$ tar -zxvf glibc-2.17.tar.gz && cd glibc-2.17
$ mkdir build && cd build
$ ../configure --prefix=/usr --with-headers=/usr/include --with-binutils=/usr/bin
$ make && make install
$ strings /lib64/libc.so.6 |grep GLIBC
GLIBC_2.2.5
GLIBC_2.2.6
GLIBC_2.3
GLIBC_2.3.2
GLIBC_2.3.3
GLIBC_2.3.4
GLIBC_2.4
GLIBC_2.5
GLIBC_2.6
GLIBC_2.7
GLIBC_2.8
GLIBC_2.9
GLIBC_2.10
GLIBC_2.11
GLIBC_2.12
GLIBC_2.13
GLIBC_2.14
GLIBC_2.15
GLIBC_2.16
GLIBC_2.17
GLIBC_PRIVATE
4 编译升级GCC到4.8.3(因为需要用到CXXABI_1.3.7，所以要求gcc版本大于4.8)
$ yum install bzip2 gcc-c++
$ wget http://ftp.gnu.org/gnu/gcc/gcc-4.8.3/gcc-4.8.3.tar.gz
$ tar -zxvf gcc-4.8.3.tar.gz
$ cd gcc-4.8.3
$ ./contrib/download_prerequisites # 脚本文件会帮我们下载、配置、安装依赖库
注：如果服务器无法连接外网，需单独下载这三个包到当前目录下，解压，并做链接；
$ wget ftp://gcc.gnu.org/pub/gcc/infrastructure/mpfr-2.4.2.tar.bz2
$ wget ftp://gcc.gnu.org/pub/gcc/infrastructure/gmp-4.3.2.tar.bz2
$ wget ftp://gcc.gnu.org/pub/gcc/infrastructure/mpc-0.8.1.tar.gz
$ tar xf mpfr-2.4.2.tar.bz2
$ tar xf gmp-4.3.2.tar.bz2
$ tar xf mpc-0.8.1.tar.gz
$ ln -s mpc-0.8.1 mpc
$ ln -s mpfr-2.4.2 mpfr
$ ln -s gmp-4.3.2 gmp
$ll gmp* mpc* mpfr* -d
lrwxrwxrwx 1 root root 9 12月 7 15:23 gmp -> gmp-4.3.2
drwxrwxrwx 15 1001 wheel 4096 1月 8 2010 gmp-4.3.2
lrwxrwxrwx 1 root root 9 12月 7 15:23 mpc -> mpc-0.8.1
drwxrwxrwx 5 1000 1000 4096 12月 8 2009 mpc-0.8.1
lrwxrwxrwx 1 root root 10 12月 7 15:23 mpfr -> mpfr-2.4.2
drwxrwxrwx 5 1114 1114 8192 11月 30 2009 mpfr-2.4.2
$ mkdir build && cd build
$ ../configure -enable-checking=release -enable-languages=c,c++ -disable-multilib
$ make && make install # 测试时make相当慢，大概走了3个小时，一般服务器30分钟
$ gcc -v # 不需要修改环境变量
使用内建 specs。
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/local/libexec/gcc/x86_64-unknown-linux-gnu/4.8.3/lto-wrapper
目标：x86_64-unknown-linux-gnu
配置为：../configure -enable-checking=release -enable-languages=c,c++ -disable-multilib
线程模型：posix
gcc 版本 4.8.3 (GCC)
$ echo -e "/usr/local/lib\n/usr/local/lib64" >/etc/ld.so.conf.d/local_libs.conf
$ ldconfig
如果报：ldconfig: /usr/local/lib64/libstdc++.so.6.0.19-gdb.py 不是 ELF 文件 - 它起始的魔数错误。
ldconfig: /usr/local/lib64/libstdc++.so.6.0.19-gdb.py is not an ELF file - it has the wrong magic bytes at the start.
$ mv /usr/local/lib64/{,bak_}libstdc++.so.6.0.19-gdb.py #改名
$ ldconfig
修改libstdc++.so.6的链接:
$ rm -f /usr/lib64/libstdc++.so.6
$ cp -a /usr/local/lib64/libstdc++.so.6.0.19 /usr/lib64/
$ ln -s /usr/lib64/libstdc++.so.6.0.19 /usr/lib64/libstdc++.so.6
$ strings /usr/lib64/libstdc++.so.6 |grep CXXABI_
CXXABI_1.3
CXXABI_1.3.1
CXXABI_1.3.2
CXXABI_1.3.3
CXXABI_1.3.4
CXXABI_1.3.5
CXXABI_1.3.6
CXXABI_1.3.7
CXXABI_TM_1
5 测试TensorFlow
$ python3 #验证
>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))
Hello, TensorFlow!
$ python3
>>> import tensorflow as tf
>>> a = tf.constant(10)
>>> b = tf.constant(32)
>>> print(sess.run(a + b))
42
$ python3
>>> import tensorflow as tf
>>> import os
>>> import shutil
>>> import numpy as np
>>> from PIL import Image
>>> import matplotlib.pyplot as plt
>>>
6 参考
http://blog.csdn.net/numen27/article/details/75332833
http://www.jianshu.com/p/fdb7b54b616e
http://blog.csdn.net/lenbow/article/details/51203526#1
https://www.tensorflow.org/install/install_linux #需翻墙
7 安装bazel 源码安装时的编译器
7.1 安装JDK1.8
google使用bazel构建tensorflow，因此我们需要编译之。首先安装64位jdk1.8，因为bazel需要java8来编译，
上传JDK1.8（jdk-8u66-linux-x64.tar.gz）安装包到/data/tools、
$ tar xf jdk-8u66-linux-x64.tar.gz
$ vim /etc/profile.d/java.sh
export JAVA_HOME=/data/tools/jdk1.8.0_66
export PATH=$JAVA_HOME/bin:$JAVA_HOME/jre/bin:$PATH
export CLASSPATH=$JAVA_HOME/lib:$JAVA_HOME/jre/lib
$ source /etc/profile.d/java_pwdx_grep.sh
$ java -version
java version "1.8.0_66"
java(TM) SE Runtime Environment (build 1.8.0_66-b17)
java HotSpot(TM) 64-Bit Server VM (build 25.66-b17, mixed mode)
7.2 编译bazel
$ git clone https://github.com/bazelbuild/bazel.git
$ cd bazel
$ git checkout -b dev 0.8.0
$ ./compile.sh
8 报错总结
8.1 找不到Glibc2.XX（ImportError: /lib64/tls/libc.so.6: version `GLIBC_2.14' not found）
glibc是GNU发布的libc库，即c运行库。 glibc是linux系统中最底层的api，几乎其它任何运行库都会依赖于glibc。 glibc除了封装linux操作系统所提供的系统服务外，它本身也提供了许多其它一些必要功能服务的实现。
由此可见，问题的根源是系统不兼容，ubuntu上用的libc 版本较高，而 CentOS 上用的版本太低导致不能执行。。
解决这个问题有三种方法：
第一种：升级Glibc，这个风险非常大，很多时候升完了发现好多东西都不能用了；
第二种：外链Glibc，也就是在其他目录建一个Glibc，然后添加一个环境变量，这个在网上看貌似是可行的，但我这么做的时候依然报错。
第三种：更换linux系统，这个问题很多时候是CentOS安装tf环境时候造成的，可以尝试更换容器
8.2 glibc: LD_LIBRARY_PATH shouldn't contain the current directory
LD_LIBRARY_PATH不能包含当前目录，需要修改环境变量并重新执行configure
echo $LD_LIBRARY_PATH # 查看
export LD_LIBRARY_PATH= # 定义
echo $LD_LIBRARY_PATH # 检查
./glibc-2.14/configure
8.3 直接升级glibc（风险比较大）
yum install gcc
wget http://ftp.gnu.org/pub/gnu/glibc/glibc-2.17.tar.xz
tar -xvf glibc-2.17.tar.xz
cd glibc-2.17
mkdir build
cd build
../configure --prefix=/usr --disable-profile --enable-add-ons --with-headers=/usr/include --with-binutils=/usr/bin
make && make install
需要等大概10分钟
8.4 外链安装glibc2
下载Glibc2.14：
http://ftp.gnu.org/gnu/glibc/或者http://www.gnu.org/software/libc/
安装：
xz -d glibc-2.14.tar.xz
tar -xvf glibc-2.14.tar
进入源码目录 建立构建目录，并cd进入构建目录：
cd glibc-2.14
mkdir build
配置：
../configure --prefix=/opt/glibc-2.14
编译安装：
make -j4
sudo make install
临时修改环境变量：
LD_LIBRARY_PATH=/opt/glibc-2.14/lib:$LD_LIBRARY_PATH
8.5 外链安装导致的严重后果
安装过程中，因为修改/etc/ld.so.conf文件，ldconfig后导致输入命令后，连最基本的命令也会报错：
ls
ls: error while loading shared libraries: __vdso_time: invalid mode for dlopen(): Invalid argument
解决方法：
千万不要断开ssh，不然就远程不上去了
vi /etc/profile 加入
export LD_LIBRARY_PATH=/usr/lib:/usr/lib64:/lib:/lib64:/usr/local/lib:/usr/local/lib64
链接完了之后，Glibc2的问题是没有了，但import tensorflow的时候出现 Segmentation fault (core dumped)
8.6 输入所有命令后都没反应了。。。
因为升级了Glibc，导致系统出问题了，把环境变量改回去就可以了。
8.7 glibc3找不到（version `GLIBCXX_3.4.21' not found）
参考http://blog.csdn.net/rznice/article/details/51090966
其实和找不到glibc2的性质差不多
strings /usr/lib64/libstdc++.so.6.0.13 |grep GLIBC
8.8 没有git
yum install git-core
要是不能联网有没有git都一样，所有包都需要手动下载
8.9 安**inutils
从以下目录下载binutils：ftp.gnu.org/gnu/binutils/binutils-2.28.tar.bz2
tar jxvf binutils-2.28.tar.bz2
mkdir binutils-build
cd binutils-build
../binutils-2.28/configure
make -j4
make install
8.10 安**azel（大坑）
下载地址1：git clone https://github.com/bazelbuild/bazel（非常之慢）
下载地址2：git clone https://github.com/CStzdong/bazel
发现报错：
INFO: You can skip this first step by providing a path to the bazel binary as second argument:
INFO: ./compile.sh compile /path/to/bazel
?? Building Bazel from scratch
ERROR: Must specify PROTOC if not bootstrapping from the distribution artifact
--------------------------------------------------------------------------------
NOTE: This failure is likely occuring if you are trying to bootstrap bazel from
a developer checkout. Those checkouts do not include the generated output of
the protoc compiler (as we prefer not to version generated files).
* To build a developer version of bazel, do
bazel build //src:bazel
* To bootstrap your first bazel binary, please download a dist archive from our
release page at https://github.com/bazelbuild/bazel/releases and run
compile.sh on the unpacked archive.
The full install instructions to install a release version of bazel can be found
at https://docs.bazel.build/install-compile-source.html
For a rationale, why the bootstrap process is organized in this way, see
https://bazel.build/designs/2016/10/11/distribution-artifact.html
进入错误信息中提到的https://github.com/bazelbuild/bazel/releases网站，选择最近版本的链接，进去后发现有一堆安装包。选择其中的一个直接下载https://github.com/bazelbuild/bazel/releases/download/0.5.3/bazel-0.5.3-installer-linux-x86_64.sh运行安装成功，执行时报错：
/usr/local/bin/bazel: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.19' not found (required by /usr/local/bin/bazel)
这个错误会在下文提到
重新运行./compile.sh
运行到一半报错
再执行一次，发现两次运行./compile.sh出现的错误不一致！疑似安装程序bug
尝试低版本bazel0.5.2，仍出现错误
尝试更低版本0.4.5，下载解压缩运行./compile.sh后安装成功！！！
下载地址：https://github.com/bazelbuild/bazel/releases/download/0.4.5/bazel-0.4.5-dist.zip
然后执行：
mkdir bazel-0.4.5-dist
cd bazel-0.4.5-dist
unzip ../bazel-0.4.5-dist.zip
./compile.sh
cp ./output/bazel /usr/local/bin（复制bazel的Binary文件至/usr/local/bin，使得全局都能找到该文件）
8.11 关于手动离线安**azel
不建议完全手动安**azel，全程有100多个包的依赖，。，，，，，，
8.12 手动安装numpy和scipy
依赖的包：
scipy-0.11.0
numpy-1.6.2
nose-1.2.1
lapack-3.4.2
atlas-3.10.0
参考：http://blog.chinaunix.net/uid-22488454-id-3978860.html
8.13 pip
如果没有pip，就到PIP官网下载get-pip.py。
参考链接：http://www.jianshu.com/p/81b648b1d572
最后从python官网下载p3安装包就好了
如果公司有自己的镜像，可以修改pip的配置文件：
cd ~/.pip/pip.conf(如果没有，就自己建一个；如果不能保存，说明没有.pip目录，需要进入~目录mkdir .pip）
然后加入下面的内容
[global]
index-url = XXX
trusted-host = pypi.douban.com
disable-pip-version-check = true
timeout = 120
注：XXX为国内或企业内部镜像，国内用https://pypi.douban.com/simple，公司内部就用自己的。
8.14 找不到readelf
依据链接http://www.jianshu.com/p/308a4e803c81的说法，先用readelf -s 文件路径|grep GLIBC_2.14查看so里到底哪部分依赖了glibc2.14，发现readelf: command not found，没有readelf命令。。。
（readelf用来显示一个或多个elf格式的目标文件信息）
依据链接http://pkgs.loginroot.com/errors/notFound/readelf，需要添加环境变量：export PATH="/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/root/bin"
8.15 Segmentation fault (core dumped)
直接强制退出Python了
根据链接https://github.com/tensorflow/tensorflow/issues/8197的解释，原因是gcc的版本过低，更新gcc在前文已经提过了。
还有文章提到是scipy和tensorflow冲突
根据http://blog.csdn.net/shouwangzhelv/article/details/51851155提到的解决方案，重新手工编译了scipy，依然不行。
8.16 安装anaconda
参考：http://www.jianshu.com/p/03d757283339
如果机器不能联网，anaconda基本就废掉了。。。
如果不能用ananconda，只好自己下载包然后上传了，单台机器就rz和sz，多台机器之间传文件就scp xxx root@abc:url
8.17 在centos系统下，导入matplotlib时，出现ImportError: No module named '_tkinter'的错误，
首先yum list installed | grep ^tk ；查看是否存在相应模块，通常原因是tkinter和tk-devel缺失。
通过yum install -y tkinter和yum install -y tk-devel下载相应模块，再重新编译Python即可。
或者编译python的时候选择添加参数--enable-unicode=ucs2
$ python3
>>> import matplotlib.pyplot as plt
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python3.5/site-packages/matplotlib/pyplot.py", line 113, in <module>
_backend_mod, new_figure_manager, draw_if_interactive, _show = pylab_setup()
File "/usr/local/lib/python3.5/site-packages/matplotlib/backends/__init__.py", line 60, in pylab_setup
[backend_name], 0)
File "/usr/local/lib/python3.5/site-packages/matplotlib/backends/backend_tkagg.py", line 6, in <module>
from six.moves import tkinter as Tk
File "/usr/local/lib/python3.5/site-packages/six.py", line 92, in __get__
result = self._resolve()
File "/usr/local/lib/python3.5/site-packages/six.py", line 115, in _resolve
return _import_module(self.mod)
File "/usr/local/lib/python3.5/site-packages/six.py", line 82, in _import_module
__import__(name)
File "/usr/local/lib/python3.5/tkinter/__init__.py", line 35, in <module>
import _tkinter # If this fails your Python may not be configured for Tk
ImportError: No module named '_tkinter'
或者参照：http://www.qttc.net/201304306.html
正确安装新版Python（在Linux中python默认是不安装Tkinter模块，）
1 首先修改Setup.dist文件
$ cd Python-3.5.4
$ cp Modules/Setup.dist{,_$(date +%F)}
$ vim Modules/Setup.dist # 把下面相应行的注释去掉，修改具体版本
_tkinter _tkinter.c tkappinit.c -DWITH_APPINIT \
-L/usr/local/lib \
-I/usr/local/include \
-ltk8.5 -ltcl8.5 \
-lX11
以上第四行-ltk8.5 -ltcl8.5 默认是 8.2 ，请你系统实际tcl/tk版本修改:我系统中装的是8.5，所以这里我改成了8.5
$ rpm -qa | grep ^tk
tk-8.5.7-5.el6.x86_64
tk-devel-8.5.7-5.el6.x86_64
tkinter-2.6.6-66.el6_8.x86_64
$ rpm -qa | grep ^tcl
tcl-8.5.7-6.el6.x86_64
tcl-devel-8.5.7-6.el6.x86_64
2 安装tck-devel、tk-devel
$ yum install tcl-devel tk-devel -y
3 开始配置安装python
$ ldconfig
$ ./configure
$ make && make install
4 验证
新版Python是否可以使用tkinter模块
$ python3
>>> import tkinter
>>>
旧版Python是否可以使用tkinter模块
$ python
>>> import Tkinter
>>>
8.18 升级gcc完，把/usr/local/lib*添加到系统动态链接库：echo -e "/usr/local/lib\n/usr/local/lib64" >/etc/ld.so.conf.d/local_lib.conf后，
执行ldconfig报错：ldconfig: /usr/local/lib64/libstdc++.so.6.0.19-gdb.py is not an ELF file - it has the wrong magic bytes at the start.
不是 ELF 文件 - 它起始的魔数错误。
9 opencv
https://m.2cto.com/kf/201610/557136.html
http://techieroop.com/install-opencv-in-centos/
http://blog.csdn.net/zl18310999566/article/details/77880862
END