갈루아의 반서재

딥러닝 학습을 위해 필요한 아나콘다, 주피터 노트북, 텐서플로우와 케라스를 설치해보자.


Downlaod Anaconda


먼저 아나콘다를 설치한다. 아나콘다 가상환경에 데이터 사이언스와 딥러닝에 필요한 패키지를 설치한다. 아나콘다를 사용하면 특정 프로젝트에 필요한 특정 버전의 패키지 설치 등 버전 충돌에 대한 우려를 덜 수 있는 장점이 있다. 다음과 같이 현재 기준으로 텐서플로우나 케라스가 지원하는 파이썬 3.6 버전을 갖춘 환경을 설치해보자.

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
founder@hilbert:~$ conda create --name tfKeras python=3.6
Solving environment: done
## Package Plan ##
  environment location: /home/founder/anaconda3/envs/tfKeras
  added / updated specs:
    - python=3.6
The following packages will be downloaded:
    package                    |            build
    ---------------------------|-----------------
    pip-18.1                   |           py36_0         1.8 MB
    setuptools-40.6.2          |           py36_0         604 KB
    python-3.6.7               |       h0371630_0        34.3 MB
    wheel-0.32.3               |           py36_0          35 KB
    certifi-2018.11.29         |           py36_0         146 KB
    ------------------------------------------------------------
                                           Total:        36.9 MB
The following NEW packages will be INSTALLED:
    ca-certificates: 2018.03.07-0
    certifi:         2018.11.29-py36_0
    libedit:         3.1.20170329-h6b74fdf_2
    libffi:          3.2.1-hd88cf55_4
    libgcc-ng:       8.2.0-hdf63c60_1
    libstdcxx-ng:    8.2.0-hdf63c60_1
    ncurses:         6.1-he6710b0_1
    openssl:         1.1.1a-h7b6447c_0
    pip:             18.1-py36_0
    python:          3.6.7-h0371630_0
    readline:        7.0-h7b6447c_5
    setuptools:      40.6.2-py36_0
    sqlite:          3.25.3-h7b6447c_0
    tk:              8.6.8-hbc83047_0
    wheel:           0.32.3-py36_0
    xz:              5.2.4-h14c3975_4
    zlib:            1.2.11-h7b6447c_3
Proceed ([y]/n)? y
Downloading and Extracting Packages
pip-18.1             | 1.8 MB    | ##################################### | 100%
setuptools-40.6.2    | 604 KB    | ##################################### | 100%
python-3.6.7         | 34.3 MB   | ##################################### | 100%
wheel-0.32.3         | 35 KB     | ################################################################################################## | 100%
certifi-2018.11.29   | 146 KB    | ################################################################################################## | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
#
# To activate this environment, use
#
#     $ conda activate tfKeras
#
# To deactivate an active environment, use
#
#     $ conda deactivate
founder@hilbert:~$ conda activate tfKeras
(tfKeras) founder@hilbert:~$
cs


Conda는 가상 환경을 관리하고 패키지를 설치해주는 패키지 매니저로, conda 를 이용하여 아래와 같은 명령을 실행할 수 있다.

# update conda in your default environment 
$ conda upgrade conda
$ conda upgrade --all
# create a new environment with conda
$ conda create -n [my-env-name]
# activate the environment you created
$ source activate [my-env-name]
# take a look at the environment you created
$ conda info
$ conda list
# install a package with conda and verify it's installed
$ conda install numpy
$ conda list
# take a look at the list of environments you currently have
$ conda info -e
# remove an environment
$ conda env remove --name [my-env-name]


특히 아나콘다는 Anaconda Cheatsheet 라는 매우 유용 문서를 제공하고 있으니 꼭 참고하길 바란다. 

Source — https://conda.io/docs/_downloads/conda-cheatsheet.pdf


Conda vs Pip install


Conda 를 사용해 생성한 가상환경에서 패키지 설치는 conda 및 pip 둘 다 사용이 가능하지만, 아래와 같은 차이점이 있다.


  • conda install  - 모든 패키지 설치 가능
  • pip install - 파이썬 패키지 설치 가능, 사실상 파이썬 패키지 매니저

pip 와 conda 는 서로 다른 패키징 포맷을 사용하므로 상호교차하여 운영할 수 없다. 그리고 conda list 를 이용하여 설치된 패키지를 확인할 수 있으며, conda 또는 pip 중 무엇을 이용하여 설치했는지도 확인할 수 있다.

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
(tfKeras) founder@hilbert:~$ conda list
# packages in environment at /home/founder/anaconda3/envs/tfKeras:
#
# Name                    Version                   Build  Channel
ca-certificates           2018.03.07                    0
certifi                   2018.11.29               py36_0
Keras                     2.2.4                     <pip>
libedit                   3.1.20170329         h6b74fdf_2
libffi                    3.2.1                hd88cf55_4
libgcc-ng                 8.2.0                hdf63c60_1
libstdcxx-ng              8.2.0                hdf63c60_1
ncurses                   6.1                  he6710b0_1
openssl                   1.1.1a               h7b6447c_0
pip                       18.1                     py36_0
python                    3.6.7                h0371630_0
PyYAML                    3.13                      <pip>
readline                  7.0                  h7b6447c_5
scipy                     1.1.0                     <pip>
setuptools                40.6.2                   py36_0
sqlite                    3.25.3               h7b6447c_0
tk                        8.6.8                hbc83047_0
wheel                     0.32.3                   py36_0
xz                        5.2.4                h14c3975_4
zlib                      1.2.11               h7b6447c_3
 
cs


Install TensorFlow (including Keras)


앞서 conda 로 생성한 가상 환경에 TenslowFlow 를 설치합니다. TensorFlow 공식 설치가이드에 따라 conda install 이 아닌 pip install 을 사용하자. 아래 링크를 참고한다. 

https://www.tensorflow.org/install/pip

각각 하나의 패키지를 설치할 수도 있고, reuqirements.txt 파일로부터 다수의 패키지를 한 번에 설치할 수도 있다.


먼저 해당 시스템의 파이썬 개발환경이 셋팅되어있는지 확인한다. python 과 pip 가 설치되어 있는지 확인하고 없다면 설치한다. 

1
2
3
4
5
(tfKeras) founder@hilbert:~$ python3 --version
Python 3.6.7 :: Anaconda, Inc.
 
(tfKeras) founder@hilbert:~$ pip3 --version
pip 9.0.1 from /usr/lib/python3/dist-packages (python 3.6)
cs

텐서플로우를 설치한다.

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
(tfKeras) founder@hilbert:~$ pip3 install --upgrade tensorflow
Collecting tensorflow
  Cache entry deserialization failed, entry ignored
  Downloading https://files.pythonhosted.org/packages/22/cc/ca70b78087015d21c5f3f93694107f34ebccb3be9624385a911d4b52ecef/tensorflow-1.12.0-cp36-cp36m-manylinux1_x86_64.whl (83.1MB)
    100|████████████████████████████████| 83.1MB 13kB/s
Collecting keras-applications>=1.0.6 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/3f/c4/2ff40221029f7098d58f8d7fb99b97e8100f3293f9856f0fb5834bef100b/Keras_Applications-1.0.6-py2.py3-none-any.whl (44kB)
    100|████████████████████████████████| 51kB 9.4MB/s
Collecting gast>=0.2.0 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/5c/78/ff794fcae2ce8aa6323e789d1f8b3b7765f601e7702726f430e814822b96/gast-0.2.0.tar.gz
Collecting six>=1.10.0 (from tensorflow)
  Cache entry deserialization failed, entry ignored
  Downloading https://files.pythonhosted.org/packages/73/fb/00a976f728d0d1fecfe898238ce23f502a721c0ac0ecfedb80e0d88c64e9/six-1.12.0-py2.py3-none-any.whl
Collecting grpcio>=1.8.6 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/3b/bb/701d879849c938028c09fdb5405dbde7c86644bbbb90098094002db23ded/grpcio-1.17.1-cp36-cp36m-manylinux1_x86_64.whl (10.1MB)
    100|████████████████████████████████| 10.1MB 131kB/s
Collecting astor>=0.6.0 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/35/6b/11530768cac581a12952a2aad00e1526b89d242d0b9f59534ef6e6a1752f/astor-0.7.1-py2.py3-none-any.whl
Collecting termcolor>=1.1.0 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/8a/48/a76be51647d0eb9f10e2a4511bf3ffb8cc1e6b14e9e4fab46173aa79f981/termcolor-1.1.0.tar.gz
Collecting absl-py>=0.1.6 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/0c/63/f505d2d4c21db849cf80bad517f0065a30be6b006b0a5637f1b95584a305/absl-py-0.6.1.tar.gz (94kB)
    100|████████████████████████████████| 102kB 10.1MB/s
Collecting numpy>=1.13.3 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/ff/7f/9d804d2348471c67a7d8b5f84f9bc59fd1cefa148986f2b74552f8573555/numpy-1.15.4-cp36-cp36m-manylinux1_x86_64.whl (13.9MB)
    100|████████████████████████████████| 13.9MB 95kB/s
Collecting protobuf>=3.6.1 (from tensorflow)
  Cache entry deserialization failed, entry ignored
  Downloading https://files.pythonhosted.org/packages/c2/f9/28787754923612ca9bfdffc588daa05580ed70698add063a5629d1a4209d/protobuf-3.6.1-cp36-cp36m-manylinux1_x86_64.whl (1.1MB)
    100|████████████████████████████████| 1.1MB 1.1MB/s
Collecting wheel>=0.26 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/ff/47/1dfa4795e24fd6f93d5d58602dd716c3f101cfd5a77cd9acbe519b44a0a9/wheel-0.32.3-py2.py3-none-any.whl
Collecting tensorboard<1.13.0,>=1.12.0 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/e0/d0/65fe48383146199f16dbd5999ef226b87bce63ad5cd73c840cf722637969/tensorboard-1.12.0-py3-none-any.whl (3.0MB)
    100|████████████████████████████████| 3.1MB 443kB/s
Collecting keras-preprocessing>=1.0.5 (from tensorflow)
  Downloading https://files.pythonhosted.org/packages/fc/94/74e0fa783d3fc07e41715973435dd051ca89c550881b3454233c39c73e69/Keras_Preprocessing-1.0.5-py2.py3-none-any.whl
Collecting h5py (from keras-applications>=1.0.6->tensorflow)
  Downloading https://files.pythonhosted.org/packages/8e/cb/726134109e7bd71d98d1fcc717ffe051767aac42ede0e7326fd1787e5d64/h5py-2.8.0-cp36-cp36m-manylinux1_x86_64.whl (2.8MB)
    100|████████████████████████████████| 2.8MB 460kB/s
Collecting setuptools (from protobuf>=3.6.1->tensorflow)
  Cache entry deserialization failed, entry ignored
  Downloading https://files.pythonhosted.org/packages/37/06/754589caf971b0d2d48f151c2586f62902d93dc908e2fd9b9b9f6aa3c9dd/setuptools-40.6.3-py2.py3-none-any.whl (573kB)
    100|████████████████████████████████| 573kB 2.0MB/s
Collecting markdown>=2.6.8 (from tensorboard<1.13.0,>=1.12.0->tensorflow)
  Downloading https://files.pythonhosted.org/packages/7a/6b/5600647404ba15545ec37d2f7f58844d690baf2f81f3a60b862e48f29287/Markdown-3.0.1-py2.py3-none-any.whl (89kB)
    100|████████████████████████████████| 92kB 10.1MB/s
Collecting werkzeug>=0.11.10 (from tensorboard<1.13.0,>=1.12.0->tensorflow)
  Downloading https://files.pythonhosted.org/packages/20/c4/12e3e56473e52375aa29c4764e70d1b8f3efa6682bef8d0aae04fe335243/Werkzeug-0.14.1-py2.py3-none-any.whl (322kB)
    100|████████████████████████████████| 327kB 3.9MB/s
Building wheels for collected packages: gast, termcolor, absl-py
  Running setup.py bdist_wheel for gast ... done
  Stored in directory: /home/founder/.cache/pip/wheels/9a/1f/0e/3cde98113222b853e98fc0a8e9924480a3e25f1b4008cedb4f
  Running setup.py bdist_wheel for termcolor ... done
  Stored in directory: /home/founder/.cache/pip/wheels/7c/06/54/bc84598ba1daf8f970247f550b175aaaee85f68b4b0c5ab2c6
  Running setup.py bdist_wheel for absl-py ... done
  Stored in directory: /home/founder/.cache/pip/wheels/18/ea/5e/e36e1b8739e78cd2eba0a08fdc602c2b16a4b263912af8cb64
Successfully built gast termcolor absl-py
Installing collected packages: six, numpy, h5py, keras-applications, gast, grpcio, astor, termcolor, absl-py, setuptools, protobuf, wheel, markdown, werkzeug, tensorboard, keras-preprocessing, tensorflow
Successfully installed absl-py-0.6.1 astor-0.7.1 gast-0.2.0 grpcio-1.17.1 h5py-2.8.0 keras-applications-1.0.6 keras-preprocessing-1.0.5 markdown-3.0.1 numpy-1.15.4 protobuf-3.6.1 setuptools-40.6.3 six-1.12.0 tensorboard-1.12.0 tensorflow-1.12.0 termcolor-1.1.0 werkzeug-0.14.1 wheel-0.32.3
 
cs

텐서플로우가 제대로 설치되었는지 확인해보자. 1.12.0 버전이 설치되었음을 확인할 수 있다.

1
2
(tfKeras) founder@hilbert:~$ python3 -'import tensorflow as tf; print(tf.__version__)'
1.12.0
cs


다음으로 케라스를 설치한다. 반드시 텐서플로우 설치 후에 진행한다.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
(tfKeras) founder@hilbert:~$ pip install Keras
Collecting Keras
  Downloading https://files.pythonhosted.org/packages/5e/10/aa32dad071ce52b5502266b5c659451cfd6ffcbf14e6c8c4f16c0ff5aaab/Keras-2.2.4-py2.py3-none-any.whl (312kB)
    100|████████████████████████████████| 317kB 21.6MB/s
Collecting pyyaml (from Keras)
  Downloading https://files.pythonhosted.org/packages/9e/a3/1d13970c3f36777c583f136c136f804d70f500168edc1edea6daa7200769/PyYAML-3.13.tar.gz (270kB)
    100|████████████████████████████████| 276kB 23.4MB/s
Requirement already satisfied: six>=1.9.0 in ./.local/lib/python3.6/site-packages (from Keras) (1.12.0)
Collecting scipy>=0.14 (from Keras)
  Downloading https://files.pythonhosted.org/packages/a8/0b/f163da98d3a01b3e0ef1cab8dd2123c34aee2bafbb1c5bffa354cc8a1730/scipy-1.1.0-cp36-cp36m-manylinux1_x86_64.whl (31.2MB)
    100|████████████████████████████████| 31.2MB 966kB/s
Requirement already satisfied: keras-preprocessing>=1.0.5 in ./.local/lib/python3.6/site-packages (from Keras) (1.0.5)
Requirement already satisfied: keras-applications>=1.0.6 in ./.local/lib/python3.6/site-packages (from Keras) (1.0.6)
Requirement already satisfied: numpy>=1.9.1 in ./.local/lib/python3.6/site-packages (from Keras) (1.15.4)
Requirement already satisfied: h5py in ./.local/lib/python3.6/site-packages (from Keras) (2.8.0)
Building wheels for collected packages: pyyaml
  Running setup.py bdist_wheel for pyyaml ... done
  Stored in directory: /home/founder/.cache/pip/wheels/ad/da/0c/74eb680767247273e2cf2723482cb9c924fe70af57c334513f
Successfully built pyyaml
Installing collected packages: pyyaml, scipy, Keras
Successfully installed Keras-2.2.4 pyyaml-3.13 scipy-1.1.0
cs

케라스 설치 여부를 확인하자. 파이썬 셀은 Ctrl+D 로 빠져나온다.

1
2
3
4
5
6
7
8
9
(tfKeras) founder@hilbert:~$ python
Python 3.6.7 |Anaconda, Inc.| (default, Oct 23 201819:16:44)
[GCC 7.3.0] on linux
Type "help""copyright""credits" or "license" for more information.
>>> import keras
Using TensorFlow backend.
kera>>> keras.__version__
'2.2.4'
 
cs


Launch Jupyter Notebook


주피터 노트북은 데이터 분석과 시각화에 필수적인 툴이다. 여기서는 원격지에 실행시킨 노트북 환경을 웹에서 구동할 있도록 다음과 같이 노트북 서버를 구동한다. 

1
2
3
4
5
6
7
8
(tfKeras) founder@hilbert:~$ jupyter notebook --no-browser --ip=0.0.0.0
[W 10:03:48.891 NotebookApp] All authentication is disabled.  Anyone who can connect to this server will be able to run code.
[I 10:03:48.922 NotebookApp] JupyterLab extension loaded from /home/founder/anaconda3/lib/python3.7/site-packages/jupyterlab
[I 10:03:48.923 NotebookApp] JupyterLab application directory is /home/founder/anaconda3/share/jupyter/lab
[I 10:03:48.926 NotebookApp] Serving notebooks from local directory: /home/founder
[I 10:03:48.927 NotebookApp] The Jupyter Notebook is running at:
[I 10:03:48.927 NotebookApp] http://(hilbert or 127.0.0.1):8888/
[I 10:03:48.927 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
cs

노트북이 가동된 모습이다.

아래는 Jupyter Notebook Cheat Sheet 이다. 다음 링크에서 다운로드 가능하다. 

https://s3.amazonaws.com/assets.datacamp.com/blog_assets/Jupyter_Notebook_Cheat_Sheet.pdf

설치된 노트북 작동 여부를 확인해보자. New > Python3 노트북을 만들고, 텐서플로우 버전을 확인하는 코드를 넣는다. 하지만 앞서 파이썬 셀에서는 정상적으로 작동하던 내용이 노트북에서는 다음과 같이 해당 모듈을 찾을 수 없다고 나온다. 

import tensorflow as tf
print ("TensorFlow version: " + tf.__version__)
---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-2-6114373f6809> in <module>()
----> 1 import tensorflow as tf
      2 print ("TensorFlow version: " + tf.__version__)

ModuleNotFoundError: No module named 'tensorflow'


이 경우 하단 링크를 참고해 에러를 해결한다. 별도 설치없이 노트북이 실행되어 이 부분 고민을 하지 않았을텐데 이 때 실행된 노트북은 기본 환경의 글로벌 주피터인셈이다. conda 환경에서 주피터를 설치해서 오버라이딩 해야한다. 

https://stackoverflow.com/questions/38221181/no-module-named-tensorflow-in-jupyter

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
(tfKeras) founder@hilbert:~$ conda install jupyter notebook
Solving environment: done
 
## Package Plan ##
 
  environment location: /home/founder/anaconda3/envs/tfKeras
 
  added / updated specs:
    - jupyter
    - notebook
 
 
The following packages will be downloaded:
 
    package                    |            build
    ---------------------------|-----------------
    ptyprocess-0.6.0           |           py36_0          23 KB
    notebook-5.7.2             |           py36_0         7.2 MB
    pyqt-5.9.2                 |   py36h05f1152_2         5.6 MB
    qt-5.9.7                   |       h5867ecd_1        85.9 MB
    pandoc-2.2.3.2             |                0        20.8 MB
    pexpect-4.6.0              |           py36_0          77 KB
    jupyter_client-5.2.3       |           py36_0         124 KB
    bleach-3.0.2               |           py36_0         220 KB
    pandocfilters-1.4.2        |           py36_1          13 KB
    wcwidth-0.1.7              |           py36_0          25 KB
    ipywidgets-7.4.2           |           py36_0         151 KB
    entrypoints-0.2.3          |           py36_2           9 KB
    markupsafe-1.1.0           |   py36h7b6447c_0          26 KB
    nbconvert-5.3.1            |           py36_0         406 KB
    pickleshare-0.7.5          |           py36_0          13 KB
    testpath-0.4.2             |           py36_0          91 KB
    terminado-0.8.1            |           py36_1          21 KB
    qtconsole-4.4.3            |           py36_0         157 KB
    decorator-4.3.0            |           py36_0          15 KB
    ipython-7.2.0              |   py36h39e3cac_0         1.0 MB
    ipython_genutils-0.2.0     |           py36_0          39 KB
    mistune-0.8.4              |   py36h7b6447c_0          54 KB
    traitlets-4.3.2            |           py36_0         133 KB
    nbformat-4.4.0             |           py36_0         141 KB
    ipykernel-5.1.0            |   py36h39e3cac_0         156 KB
    prompt_toolkit-2.0.7       |           py36_0         482 KB
    tornado-5.1.1              |   py36h7b6447c_0         663 KB
    python-dateutil-2.7.5      |           py36_0         275 KB
    six-1.12.0                 |           py36_0          22 KB
    jupyter-1.0.0              |           py36_7           6 KB
    parso-0.3.1                |           py36_0         114 KB
    pyzmq-17.1.2               |   py36h14c3975_0         454 KB
    jupyter_core-4.4.0         |           py36_0          63 KB
    pygments-2.2.0             |           py36_0         1.3 MB
    jedi-0.13.1                |           py36_0         229 KB
    widgetsnbextension-3.4.2   |           py36_0         1.7 MB
    prometheus_client-0.4.2    |           py36_0          64 KB
    sip-4.19.8                 |   py36hf484d3e_0         290 KB
    webencodings-0.5.1         |           py36_1          19 KB
    jupyter_console-6.0.0      |           py36_0          35 KB
    jsonschema-2.6.0           |           py36_0          62 KB
    libpng-1.6.35              |       hbc83047_0         335 KB
    jinja2-2.10                |           py36_0         184 KB
    backcall-0.1.0             |           py36_0          19 KB
    send2trash-1.5.0           |           py36_0          16 KB
    ------------------------------------------------------------
                                           Total:       128.6 MB
 
The following NEW packages will be INSTALLED:
 
    backcall:           0.1.0-py36_0
    bleach:             3.0.2-py36_0
    dbus:               1.13.2-h714fa37_1
    decorator:          4.3.0-py36_0
    entrypoints:        0.2.3-py36_2
    expat:              2.2.6-he6710b0_0
    fontconfig:         2.13.0-h9420a91_0
    freetype:           2.9.1-h8a8886c_1
    glib:               2.56.2-hd408876_0
    gmp:                6.1.2-h6c8ec71_1
    gst-plugins-base:   1.14.0-hbbd80ab_1
    gstreamer:          1.14.0-hb453b48_1
    icu:                58.2-h9c2bf20_1
    ipykernel:          5.1.0-py36h39e3cac_0
    ipython:            7.2.0-py36h39e3cac_0
    ipython_genutils:   0.2.0-py36_0
    ipywidgets:         7.4.2-py36_0
    jedi:               0.13.1-py36_0
    jinja2:             2.10-py36_0
    jpeg:               9b-h024ee3a_2
    jsonschema:         2.6.0-py36_0
    jupyter:            1.0.0-py36_7
    jupyter_client:     5.2.3-py36_0
    jupyter_console:    6.0.0-py36_0
    jupyter_core:       4.4.0-py36_0
    libpng:             1.6.35-hbc83047_0
    libsodium:          1.0.16-h1bed415_0
    libuuid:            1.0.3-h1bed415_2
    libxcb:             1.13-h1bed415_1
    libxml2:            2.9.8-h26e45fe_1
    markupsafe:         1.1.0-py36h7b6447c_0
    mistune:            0.8.4-py36h7b6447c_0
    nbconvert:          5.3.1-py36_0
    nbformat:           4.4.0-py36_0
    notebook:           5.7.2-py36_0
    pandoc:             2.2.3.2-0
    pandocfilters:      1.4.2-py36_1
    parso:              0.3.1-py36_0
    pcre:               8.42-h439df22_0
    pexpect:            4.6.0-py36_0
    pickleshare:        0.7.5-py36_0
    prometheus_client:  0.4.2-py36_0
    prompt_toolkit:     2.0.7-py36_0
    ptyprocess:         0.6.0-py36_0
    pygments:           2.2.0-py36_0
    pyqt:               5.9.2-py36h05f1152_2
    python-dateutil:    2.7.5-py36_0
    pyzmq:              17.1.2-py36h14c3975_0
    qt:                 5.9.7-h5867ecd_1
    qtconsole:          4.4.3-py36_0
    send2trash:         1.5.0-py36_0
    sip:                4.19.8-py36hf484d3e_0
    six:                1.12.0-py36_0
    terminado:          0.8.1-py36_1
    testpath:           0.4.2-py36_0
    tornado:            5.1.1-py36h7b6447c_0
    traitlets:          4.3.2-py36_0
    wcwidth:            0.1.7-py36_0
    webencodings:       0.5.1-py36_1
    widgetsnbextension: 3.4.2-py36_0
    zeromq:             4.2.5-hf484d3e_1
 
Proceed ([y]/n)? y
 
 
Downloading and Extracting Packages
ptyprocess-0.6.0     | 23 KB     | ################################################################################################## | 100%
notebook-5.7.2       | 7.2 MB    | ################################################################################################## | 100%
pyqt-5.9.2           | 5.6 MB    | ################################################################################################## | 100%
qt-5.9.7             | 85.9 MB   | ################################################################################################## | 100%
pandoc-2.2.3.2       | 20.8 MB   | ################################################################################################## | 100%
pexpect-4.6.0        | 77 KB     | ################################################################################################## | 100%
jupyter_client-5.2.3 | 124 KB    | ################################################################################################## | 100%
bleach-3.0.2         | 220 KB    | ################################################################################################## | 100%
pandocfilters-1.4.2  | 13 KB     | ################################################################################################## | 100%
wcwidth-0.1.7        | 25 KB     | ################################################################################################## | 100%
ipywidgets-7.4.2     | 151 KB    | ################################################################################################## | 100%
entrypoints-0.2.3    | 9 KB      | ################################################################################################## | 100%
markupsafe-1.1.0     | 26 KB     | ################################################################################################## | 100%
nbconvert-5.3.1      | 406 KB    | ################################################################################################## | 100%
pickleshare-0.7.5    | 13 KB     | ################################################################################################## | 100%
testpath-0.4.2       | 91 KB     | ################################################################################################## | 100%
terminado-0.8.1      | 21 KB     | ################################################################################################## | 100%
qtconsole-4.4.3      | 157 KB    | ################################################################################################## | 100%
decorator-4.3.0      | 15 KB     | ################################################################################################## | 100%
ipython-7.2.0        | 1.0 MB    | ################################################################################################## | 100%
ipython_genutils-0.2 | 39 KB     | ################################################################################################## | 100%
mistune-0.8.4        | 54 KB     | ################################################################################################## | 100%
traitlets-4.3.2      | 133 KB    | ################################################################################################## | 100%
nbformat-4.4.0       | 141 KB    | ################################################################################################## | 100%
ipykernel-5.1.0      | 156 KB    | ################################################################################################## | 100%
prompt_toolkit-2.0.7 | 482 KB    | ################################################################################################## | 100%
tornado-5.1.1        | 663 KB    | ################################################################################################## | 100%
python-dateutil-2.7. | 275 KB    | ################################################################################################## | 100%
six-1.12.0           | 22 KB     | ################################################################################################## | 100%
jupyter-1.0.0        | 6 KB      | ################################################################################################## | 100%
parso-0.3.1          | 114 KB    | ################################################################################################## | 100%
pyzmq-17.1.2         | 454 KB    | ################################################################################################## | 100%
jupyter_core-4.4.0   | 63 KB     | ################################################################################################## | 100%
pygments-2.2.0       | 1.3 MB    | ################################################################################################## | 100%
jedi-0.13.1          | 229 KB    | ################################################################################################## | 100%
widgetsnbextension-3 | 1.7 MB    | ################################################################################################## | 100%
prometheus_client-0. | 64 KB     | ################################################################################################## | 100%
sip-4.19.8           | 290 KB    | ################################################################################################## | 100%
webencodings-0.5.1   | 19 KB     | ################################################################################################## | 100%
jupyter_console-6.0. | 35 KB     | ################################################################################################## | 100%
jsonschema-2.6.0     | 62 KB     | ################################################################################################## | 100%
libpng-1.6.35        | 335 KB    | ################################################################################################## | 100%
jinja2-2.10          | 184 KB    | ################################################################################################## | 100%
backcall-0.1.0       | 19 KB     | ################################################################################################## | 100%
send2trash-1.5.0     | 16 KB     | ################################################################################################## | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
 
cs

conda list 를 실행해보면 패키지 목록에 주피터가 들어간 것을 확인할 수 있다. 앞서 실행했을 때는 이 패키지가 존재하지 않았다.

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
(tfKeras) founder@hilbert:~$ conda list
# packages in environment at /home/founder/anaconda3/envs/tfKeras:
#
# Name                    Version                   Build  Channel
backcall                  0.1.0                    py36_0
bleach                    3.0.2                    py36_0
ca-certificates           2018.03.07                    0
certifi                   2018.11.29               py36_0
dbus                      1.13.2               h714fa37_1
decorator                 4.3.0                    py36_0
entrypoints               0.2.3                    py36_2
expat                     2.2.6                he6710b0_0
fontconfig                2.13.0               h9420a91_0
freetype                  2.9.1                h8a8886c_1
glib                      2.56.2               hd408876_0
gmp                       6.1.2                h6c8ec71_1
gst-plugins-base          1.14.0               hbbd80ab_1
gstreamer                 1.14.0               hb453b48_1
icu                       58.2                 h9c2bf20_1
ipykernel                 5.1.0            py36h39e3cac_0
ipython                   7.2.0            py36h39e3cac_0
ipython_genutils          0.2.0                    py36_0
ipywidgets                7.4.2                    py36_0
jedi                      0.13.1                   py36_0
jinja2                    2.10                     py36_0
jpeg                      9b                   h024ee3a_2
jsonschema                2.6.0                    py36_0
jupyter                   1.0.0                    py36_7
jupyter_client            5.2.3                    py36_0
jupyter_console           6.0.0                    py36_0
jupyter_core              4.4.0                    py36_0
Keras                     2.2.4                     <pip>
libedit                   3.1.20170329         h6b74fdf_2
libffi                    3.2.1                hd88cf55_4
libgcc-ng                 8.2.0                hdf63c60_1
libpng                    1.6.35               hbc83047_0
libsodium                 1.0.16               h1bed415_0
libstdcxx-ng              8.2.0                hdf63c60_1
libuuid                   1.0.3                h1bed415_2
libxcb                    1.13                 h1bed415_1
libxml2                   2.9.8                h26e45fe_1
markupsafe                1.1.0            py36h7b6447c_0
mistune                   0.8.4            py36h7b6447c_0
nbconvert                 5.3.1                    py36_0
nbformat                  4.4.0                    py36_0
ncurses                   6.1                  he6710b0_1
notebook                  5.7.2                    py36_0
openssl                   1.1.1a               h7b6447c_0
pandoc                    2.2.3.2                       0
pandocfilters             1.4.2                    py36_1
parso                     0.3.1                    py36_0
pcre                      8.42                 h439df22_0
pexpect                   4.6.0                    py36_0
pickleshare               0.7.5                    py36_0
pip                       18.1                     py36_0
prometheus_client         0.4.2                    py36_0
prompt_toolkit            2.0.7                    py36_0
ptyprocess                0.6.0                    py36_0
pygments                  2.2.0                    py36_0
pyqt                      5.9.2            py36h05f1152_2
python                    3.6.7                h0371630_0
python-dateutil           2.7.5                    py36_0
PyYAML                    3.13                      <pip>
pyzmq                     17.1.2           py36h14c3975_0
qt                        5.9.7                h5867ecd_1
qtconsole                 4.4.3                    py36_0
readline                  7.0                  h7b6447c_5
scipy                     1.1.0                     <pip>
send2trash                1.5.0                    py36_0
setuptools                40.6.2                   py36_0
sip                       4.19.8           py36hf484d3e_0
six                       1.12.0                   py36_0
sqlite                    3.25.3               h7b6447c_0
terminado                 0.8.1                    py36_1
testpath                  0.4.2                    py36_0
tk                        8.6.8                hbc83047_0
tornado                   5.1.1            py36h7b6447c_0
traitlets                 4.3.2                    py36_0
wcwidth                   0.1.7                    py36_0
webencodings              0.5.1                    py36_1
wheel                     0.32.3                   py36_0
widgetsnbextension        3.4.2                    py36_0
xz                        5.2.4                h14c3975_4
zeromq                    4.2.5                hf484d3e_1
zlib                      1.2.11               h7b6447c_3
 
cs

다시 위 코드를 실행해보자. 정상적으로 출력된다.

import tensorflow as tf
print ("TensorFlow version: " + tf.__version__)
TensorFlow version: 1.12.0


원문출처 https://medium.com/@margaretmz/anaconda-jupyter-notebook-tensorflow-and-keras-b91f381405f8