Tensor Flow
環境：
Core i7-7700HQ 定格2.80GHz
GTX1070 8GB
8GB DDR4 SO-DIMM
Win10 Home 64 bit
まずはインストール
Installing TensorFlow on Windows | TensorFlow
1. CUDA
Installation Guide Windows :: CUDA Toolkit Documentation
Patch2もインストール。
2．cuDNN
NVIDIA cuDNN | NVIDIA Developer
3. Anaconda
Download Anaconda Now! | Continuum
最新4.4をインストール
Python3.6なのだが、環境を作るときにpython3.5を指定する。
Anaconda Prompt起動
C:> conda create -n tensorflow python=3.5
C:> activate tensorflow
# Install GPU version tensor flow
(tensorflow)C:> pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/windows/gpu/tensorflow_gpu-1.2.1-cp35-cp35m-win_amd64.whl
５．Validate
>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))一行ずつ入力・・・
>>> import tensorflow as tfエラー・・・
ImportError: No module named '_pywrap_tensorflow_internal'
検索・・・
windowsにネイティブサポートしたTensorflowを動かす - Qiita
なるほど、VS Redistributableが足りないらしい？
入れてみて、もう一度！
ちがうみたい。
Redistは、anacondaでインストール済みのようだ。
こっちの記事
CUDA - TensorFlow importエラーについて(74099)｜teratail
cuDNN 6.0 -> 5.1にインストールしなおし。。。
ファイルを置き換え
>>> import tensorflow as tf成功！
>>> sess = tf.Session()
2017-07-22 02:27:49.195206: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.195814: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE2 instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.197318: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.197765: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.198217: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.198701: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.199164: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:49.199214: W c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\platform\cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
2017-07-22 02:27:50.630526: I c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\common_runtime\gpu\gpu_device.cc:940] Found device 0 with properties:
name: GeForce GTX 1070
major: 6 minor: 1 memoryClockRate (GHz) 1.645
pciBusID 0000:01:00.0
Total memory: 8.00GiB
Free memory: 6.67GiB
2017-07-22 02:27:50.630665: I c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\common_runtime\gpu\gpu_device.cc:961] DMA: 0
2017-07-22 02:27:50.631359: I c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\common_runtime\gpu\gpu_device.cc:971] 0: Y
2017-07-22 02:27:50.631775: I c:\tf_jenkins\home\workspace\release-win\m\windows-gpu\py\35\tensorflow\core\common_runtime\gpu\gpu_device.cc:1030] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)
なんかいっぱいログが出た。
>>> print(sess.run(hello))
b'Hello, TensorFlow!'
よし！なんとか完了。
とりあえずここまで。
次回は
Getting Started With TensorFlow | TensorFlow
にすすむ
