UbuntuにTensorFlowを導入してみる

2016/04/10 2016/12/03

TensorFlowとは

Google製のディープラーニングを試せるライブラリです。

※個人的なメモです。CPUアーキテクチャの種類、PythonのバージョンやGPUの利用によっては手順が変わってくるので基本は公式のドキュメントを参照したほうが良いです。

実施環境

x64 アーキテクチャ
Ubuntu Trusty (14.04 LTS)
GPU無し

導入手順

pipのインストール

sudo apt-get install python-pip python-dev

1	sudo apt-get install python-pip python-dev

tensorflowのインストール(CPU python2.7版)

sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.7.1-cp27-none-linux_x86_64.whl

1	sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.7.1-cp27-none-linux_x86_64.whl

ディレクトリの確認

python -c 'import os; import inspect; import tensorflow; print(os.path.dirname(inspect.getfile(tensorflow)))'

1	python -c 'import os; import inspect; import tensorflow; print(os.path.dirname(inspect.getfile(tensorflow)))'

mnistの実行

python -m tensorflow.models.image.mnist.convolutional

1	python -m tensorflow.models.image.mnist.convolutional

実行結果

Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.
Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.
Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.
Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.
Extracting data/train-images-idx3-ubyte.gz
Extracting data/train-labels-idx1-ubyte.gz
Extracting data/t10k-images-idx3-ubyte.gz
Extracting data/t10k-labels-idx1-ubyte.gz
Initialized!
Step 0 (epoch 0.00), 7.0 ms
Minibatch loss: 12.053, learning rate: 0.010000
Minibatch error: 90.6%
Validation error: 84.6%

...

Step 8500 (epoch 9.89), 569.6 ms
Minibatch loss: 1.601, learning rate: 0.006302
Minibatch error: 0.0%
Validation error: 0.9%
Test error: 0.8%

Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.

Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.

Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.

Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.

Extracting data/train-images-idx3-ubyte.gz

Extracting data/train-labels-idx1-ubyte.gz

Extracting data/t10k-images-idx3-ubyte.gz

Extracting data/t10k-labels-idx1-ubyte.gz

Initialized!

Step 0 (epoch 0.00), 7.0 ms

Minibatch loss: 12.053, learning rate: 0.010000

Minibatch error: 90.6%

Validation error: 84.6%

...

Step 8500 (epoch 9.89), 569.6 ms

Minibatch loss: 1.601, learning rate: 0.006302

Minibatch error: 0.0%

Validation error: 0.9%

Test error: 0.8%

gitのインストール

sudo apt-get install git

1	sudo apt-get install git

TensorFlow repositoryのclone

git clone --recurse-submodules https://github.com/tensorflow/tensorflow

1	git clone --recurse-submodules https://github.com/tensorflow/tensorflow

JDK8のインストール

sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update
sudo apt-get install oracle-java8-installer

sudo add-apt-repository ppa:webupd8team/java

sudo apt-get update

sudo apt-get install oracle-java8-installer

その他Bazelのインストールに必要なパッケージの導入

sudo apt-get install pkg-config zip g++ zlib1g-dev unzip

1	sudo apt-get install pkg-config zip g++ zlib1g-dev unzip

Bazelのインストール

wget https://github.com/bazelbuild/bazel/releases/download/0.2.1/bazel-0.2.1-installer-linux-x86_64.sh
chmod +x bazel-0.2.1-installer-linux-x86_64.sh
./bazel-0.2.1-installer-linux-x86_64.sh --user

wget https://github.com/bazelbuild/bazel/releases/download/0.2.1/bazel-0.2.1-installer-linux-x86_64.sh

chmod +x bazel-0.2.1-installer-linux-x86_64.sh

./bazel-0.2.1-installer-linux-x86_64.sh --user

環境設定

echo "export PATH=\"\$PATH:\$HOME/bin\"" >> .bashrc
source .bashrc

1 2	echo "export PATH=\"\$PATH:\$HOME/bin\"" >> .bashrc source .bashrc

その他必要資材のインストール

sudo apt-get install python-numpy swig python-dev

1	sudo apt-get install python-numpy swig python-dev

Tensorflowのconfigure

cd ~/tensorflow
./configure
Please specify the location of python. [Default is /usr/bin/python]:
Do you wish to build TensorFlow with GPU support? [y/N] N
No GPU support will be enabled for TensorFlow
Configuration finished

cd ~/tensorflow

./configure

Please specify the location of python. [Default is /usr/bin/python]:

Do you wish to build TensorFlow with GPU support? [y/N] N

No GPU support will be enabled for TensorFlow

Configuration finished

Tensorflowの設定

cd ~/tensorflow/
bazel build -c opt //tensorflow/tools/pip_package:build_pip_package
mkdir _python_build
cd _python_build
ln -s ../bazel-bin/tensorflow/tools/pip_package/build_pip_package.runfiles/* .
ln -s ../tensorflow/tools/pip_package/* .
sudo python setup.py develop

cd ~/tensorflow/

bazel build -c opt //tensorflow/tools/pip_package:build_pip_package

mkdir _python_build

cd _python_build

ln -s ../bazel-bin/tensorflow/tools/pip_package/build_pip_package.runfiles/* .

ln -s ../tensorflow/tools/pip_package/* .

sudo python setup.py develop

bazel buildでエラーが発生したverbose_failuresオプションでエラー箇所がわかるようなのでオプションをつけて再実行

ERROR: /home/dorako321/tensorflow/tensorflow/core/kernels/BUILD:256:1: C++ compilation of rule '//tensorflow/core/kernels:mirror_pad_op' failed: gcc failed: error executing command /usr/bin/gcc -U_FORTIFY_SOURCE '-D_FORTIFY_SOURCE=1' -fstack-protector -Wall -Wunused-but-set-parameter -Wno-free-nonheap-object -fno-omit-frame-pointer -g0 -O2 -DNDEBUG -ffunction-sections ... (remaining 70 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 4.
In file included from external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/Tensor:82:0,
                 from ./third_party/eigen3/unsupported/Eigen/CXX11/Tensor:1,
                 from ./tensorflow/core/kernels/mirror_pad_op.h:19,
                 from tensorflow/core/kernels/mirror_pad_op.cc:20:
external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h: In constructor 'Eigen::internal::TensorUInt128<HIGH, LOW>::TensorUInt128(const T&)':
external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h:56:76: warning: typedef 'UnsignedT' locally defined but not used [-Wunused-local-typedefs]
     typedef typename conditional<sizeof(T) == 8, uint64_t, uint32_t>::type UnsignedT;
                                                                            ^
external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h:57:78: warning: typedef 'UnsignedLow' locally defined but not used [-Wunused-local-typedefs]
     typedef typename conditional<sizeof(LOW) == 8, uint64_t, uint32_t>::type UnsignedLow;
                                                                              ^
gcc: internal compiler error: Killed (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <file:///usr/share/doc/gcc-4.8/README.Bugs> for instructions.
Target //tensorflow/tools/pip_package:build_pip_package failed to build
Use --verbose_failures to see the command lines of failed build steps.

ERROR: /home/dorako321/tensorflow/tensorflow/core/kernels/BUILD:256:1: C++ compilation of rule '//tensorflow/core/kernels:mirror_pad_op' failed: gcc failed: error executing command /usr/bin/gcc -U_FORTIFY_SOURCE '-D_FORTIFY_SOURCE=1' -fstack-protector -Wall -Wunused-but-set-parameter -Wno-free-nonheap-object -fno-omit-frame-pointer -g0 -O2 -DNDEBUG -ffunction-sections ... (remaining 70 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 4.

In file included from external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/Tensor:82:0,

from ./third_party/eigen3/unsupported/Eigen/CXX11/Tensor:1,

from ./tensorflow/core/kernels/mirror_pad_op.h:19,

from tensorflow/core/kernels/mirror_pad_op.cc:20:

external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h: In constructor 'Eigen::internal::TensorUInt128<HIGH, LOW>::TensorUInt128(const T&)':

external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h:56:76: warning: typedef 'UnsignedT' locally defined but not used [-Wunused-local-typedefs]

typedef typename conditional<sizeof(T) == 8, uint64_t, uint32_t>::type UnsignedT;

external/eigen_archive/eigen-eigen-6e521c802bf5/unsupported/Eigen/CXX11/src/Tensor/TensorUInt128.h:57:78: warning: typedef 'UnsignedLow' locally defined but not used [-Wunused-local-typedefs]

typedef typename conditional<sizeof(LOW) == 8, uint64_t, uint32_t>::type UnsignedLow;

gcc: internal compiler error: Killed (program cc1plus)

Please submit a full bug report,

with preprocessed source if appropriate.

See <file:///usr/share/doc/gcc-4.8/README.Bugs> for instructions.

Target //tensorflow/tools/pip_package:build_pip_package failed to build

Use --verbose_failures to see the command lines of failed build steps.

verboseオプション付き

cd ~/tensorflow/
bazel build -c opt //tensorflow/tools/pip_package:build_pip_package --verbose_failures

1 2	cd ~/tensorflow/ bazel build -c opt //tensorflow/tools/pip_package:build_pip_package --verbose_failures

検索してみると、どうやら単純にメモリ不足でエラーとなっている模様。というわけでVMの割当メモリを増やして再度ビルド。

TensorFlowのニューラルネットモデルの学習

cd ~/tensorflow/tensorflow/models/image/mnist
python convolutional.py

Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.
Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.
Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.
Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.
Extracting data/train-images-idx3-ubyte.gz
Extracting data/train-labels-idx1-ubyte.gz
Extracting data/t10k-images-idx3-ubyte.gz
Extracting data/t10k-labels-idx1-ubyte.gz
Initialized!
Step 0 (epoch 0.00), 5.8 ms
Minibatch loss: 12.053, learning rate: 0.010000
Minibatch error: 90.6%
Validation error: 84.6%


...

Step 8500 (epoch 9.89), 524.5 ms
Minibatch loss: 1.594, learning rate: 0.006302
Minibatch error: 0.0%
Validation error: 0.8%
Test error: 0.8%

cd ~/tensorflow/tensorflow/models/image/mnist

python convolutional.py