Start studying TensorFlow.
First, prepare the environment. The platform is OS X El Capitan (10.11.6).
I refer to the official documentation Download and Setup.
Prepare the Python in the Requirements of Download and Setup (https://www.tensorflow.org/versions/r0.11/get_started/os_setup.html#install-ipython). Python is originally installed on OS X (macOS), but I will install it separately to use the 3rd system. I decided to use pyenv to install Python.
I originally used anyenv, so I used ʻanyenv to install pyenvand thenpyenv` to install python 3.5.2. ..
% anyenv install pyenv
Cloning https://github.com/yyuu/pyenv.git...
Cloning into 'pyenv'...
remote: Counting objects: 14083, done.
remote: Compressing objects: 100% (7/7), done.
remote: Total 14083 (delta 0), reused 0 (delta 0), pack-reused 14075
Receiving objects: 100% (14083/14083), 2.48 MiB | 1.13 MiB/s, done.
Resolving deltas: 100% (9738/9738), done.
Checking connectivity... done.
Install pyenv succeeded!
Please reload your profile (exec $SHELL -l) or open a new session.
% exec $SHELL -l
% pyenv install 3.5.2
Downloading Python-3.5.2.tar.xz...
-> https://www.python.org/ftp/python/3.5.2/Python-3.5.2.tar.xz
Installing Python-3.5.2...
patching file Lib/venv/scripts/posix/activate.fish
You can always use Python 3.5.2 in this environment, so specify the version of Python to use with pyenv global.
% pyenv global 3.5.2
% exec $SHELL -l
% python -V
Python 3.5.2
Requirements describes the dependency when using the GPU version, but I will not use it for the time being.
In Overview, consider the range of impact on existing Python-based applications when installing TensorFlow. And some options are shown. This time, I have already installed a dedicated Python environment using pyenv, so I don't have to worry about the impact on the existing environment, so it's easy Pip Installation Select .11 / get_started / os_setup.html # pip-installation).
pip is already installed, but for the time being, use ʻeasy_install` according to the procedure in the official documentation.
% easy_install pip
% easy_install --upgrade six
Next, set the appropriate URL for each environment in the environment variable TF_BINARY_URL. This time, I chose "Mac OS X, CPU only, Python 3.4 or 3.5".
% export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.11.0rc2-py3-none-any.whl
% pip3 install --upgrade $TF_BINARY_URL
Collecting tensorflow==0.11.0rc2 from https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.11.0rc2-py3-none-any.whl
Downloading https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.11.0rc2-py3-none-any.whl (35.5MB)
100% |████████████████████████████████| 35.5MB 43kB/s
Collecting numpy>=1.11.0 (from tensorflow==0.11.0rc2)
Downloading numpy-1.11.2-cp35-cp35m-macosx_10_6_intel.macosx_10_9_intel.macosx_10_9_x86_64.macosx_10_10_intel.macosx_10_10_x86_64.whl (3.9MB)
100% |████████████████████████████████| 3.9MB 391kB/s
Collecting protobuf==3.0.0 (from tensorflow==0.11.0rc2)
Downloading protobuf-3.0.0-py2.py3-none-any.whl (342kB)
100% |████████████████████████████████| 348kB 3.3MB/s
Requirement already up-to-date: six>=1.10.0 in /Users/xxxxxx/.anyenv/envs/pyenv/versions/3.5.2/lib/python3.5/site-packages/six-1.10.0-py3.5.egg (from tensorflow==0.11.0rc2)
Collecting wheel>=0.26 (from tensorflow==0.11.0rc2)
Downloading wheel-0.29.0-py2.py3-none-any.whl (66kB)
100% |████████████████████████████████| 71kB 5.8MB/s
Collecting setuptools (from protobuf==3.0.0->tensorflow==0.11.0rc2)
Downloading setuptools-28.7.1-py2.py3-none-any.whl (472kB)
100% |████████████████████████████████| 481kB 2.3MB/s
Installing collected packages: numpy, setuptools, protobuf, wheel, tensorflow
Found existing installation: setuptools 20.10.1
Uninstalling setuptools-20.10.1:
Successfully uninstalled setuptools-20.10.1
Successfully installed numpy-1.11.2 protobuf-3.0.0 setuptools-28.7.1 tensorflow-0.11.0rc2 wheel-0.29.0
You are using pip version 8.1.1, however version 9.0.0 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.
Looking at this, six is also included in the dependency, so it may not have been necessary to upgrade in advance.
And finally, there was a message prompting me to update the version of pip, so I updated it.
% pip install --upgrade pip
Collecting pip
Downloading pip-9.0.0-py2.py3-none-any.whl (1.3MB)
100% |████████████████████████████████| 1.3MB 1.0MB/s
Installing collected packages: pip
Found existing installation: pip 8.1.1
Uninstalling pip-8.1.1:
Successfully uninstalled pip-8.1.1
Successfully installed pip-9.0.0
Check the operation by the method of Run a TensorFlow demo model.
% python
Python 3.5.2 (default, Nov 3 2016, 18:42:02)
[GCC 4.2.1 Compatible Apple LLVM 6.1.0 (clang-602.0.53)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))
b'Hello, TensorFlow!'
>>> a = tf.constant(10)
>>> b = tf.constant(32)
>>> print(sess.run(a + b))
42
The display of'Hello, TensorFlow!'Is a little different from the documentation, but I think this is because the display of b'string' has changed between Python 2.7 and 3.x.
Now you're ready to go.
Recommended Posts