Deploying Deep Learning development environment on Ubuntu 16.04

Bing 20 Jul 2017

System Requirements

The GPU-enabled version of TensorFlow and Theano has the following requirements:

Ubuntu 16.04
GTX 1080
Python 2.7
CUDA 8.0
cuDNN v5.1
TensorFlow v1.2
Theano v0.9
gpuarray v0.6

Pre-deploy

Update system

sudo apt-get update && sudo apt-get -y upgrade

Install essential build tools for system

sudo apt-get install -y build-essential &&\
sudo apt-get install -y git &&\
sudo apt-get install -y python-pip &&\
sudo apt-get install -y python-matplotlib &&\
sudo apt-get install -y python-dev &&\
sudo apt-get install -y python-pydot &&\
sudo apt-get install -y python-numpy &&\
sudo apt-get install -y python-sklearn &&\
sudo apt-get install -y python-pandas &&\
sudo apt-get install -y libfreetype6-dev &&\
sudo apt-get install -y libxft-dev &&\
sudo apt-get install -y libncurses-dev &&\
sudo apt-get install -y libopenblas-dev &&\
sudo apt-get install -y gfortran &&\
sudo apt-get install -y libblas-dev &&\
sudo apt-get install -y liblapack-dev &&\
sudo apt-get install -y libatlas-base-dev &&\
sudo apt-get install -y unzip zip &&\
sudo apt-get install -y swig python-pandas &&\
sudo apt-get install -y wget &&\
sudo apt-get install -y pkg-config &&\
sudo apt-get install -y zlib1g-dev

Install CUDA 8.0 for GTX 1080

Verify You Have a CUDA-Capable GPU
```
lspci | grep -i nvidia
```
If you do not see any settings, update the PCI hardware database that Linux maintains by entering update-pciids (generally found in /sbin) at the command line and rerun the previous lspci command.
Verify the System has the Correct Kernel Headers and Development Packages Installed

The CUDA Driver requires that the kernel headers and development packages for the running version of the kernel be installed at the time of the driver installation, as well whenever the driver is rebuilt
```
sudo apt-get install linux-headers-$(uname -r)
```
Download the NVIDIA CUDA Toolkit

The NVIDIA CUDA Toolkit is available at NVIDIA Download.

Choose the platform you are using and download the NVIDIA CUDA Toolkit. The CUDA Toolkit contains the CUDA driver and tools needed to create, build and run a CUDA application as well as libraries, header files, CUDA samples source code, and other resources.

Install CUDA

sudo dpkg -i cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
sudo apt-get update
sudo apt-get install cuda

Configure environment path

system library (as admin)

Append the following contents to the “/etc/profile”

export CUDA_HOME=/usr/local/cuda
export CUDA_ROOT=/usr/local/cuda
export PATH=${CUDA_HOME}/bin:$PATH
export LD_LIBRARY_PATH=${CUDA_HOME}/lib64:$LD_LIBRARY_PATH

Make change effect and Verify it.

source /etc/profile
echo ${CUDA_HOME}

user library

Append the following contents to the “~/.bashrc”

export CUDA_HOME=/usr/local/cuda
export CUDA_ROOT=/usr/local/cuda
export PATH=${CUDA_HOME}/bin:$PATH
export LD_LIBRARY_PATH=${CUDA_HOME}/lib64:$LD_LIBRARY_PATH

Make change effect and Verify it.

source ~/.bashrc
echo ${CUDA_HOME}

Verify a Successful Installation

cd ${CUDA_HOME}/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery

Install cuDNN v5.1

Once the CUDA Toolkit is installed, download cuDNN v5.1 Library for Linux and install by following the official documentation.

Once downloaded, navigate to the directory containing cuDNN then execute following command:

tar xf cudnn-8.0-linux-x64-v5.1.tgz -C /usr/local

Install TensorFlow v1.2

Prepare TensorFlow Dependencies and Required Packages
```
sudo apt-get install libcupti-dev
```

Install TensorFlow (GPU-accelerated version)

sudo pip  install --upgrade https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-1.2.1-cp27-none-linux_x86_64.whl

Verify a Successful Installation

Invoke python: type python in command line

python

Input the following short program:

import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!')
sess = tf.Session()
print(sess.run(hello))

Install theano

Please refer to Theano install Documents to satisfy the requirements.
Install the latest stable version of Theano with:
```
sudo pip install Theano[test, doc]
```

Install gpuarray

Please double check Build ad Run Requirements based on Installation Document

Download source code.

git clone https://github.com/Theano/libgpuarray.git
cd libgpuarray

Step-by-step install: system library (as admin)

For libgpuarray:

mkdir Build
cd Build
cmake .. -DCMAKE_BUILD_TYPE=Release
make
make install
cd ..

For pygpu:

# This must be done after libgpuarray is installed as per instructions above.
python setup.py build
python setup.py install

Reference

Stats:

0 comments

Prev: Install Tesseract 4.0 in Ubuntu 16.04 Next： Robust Contextual Transformer Network for Inertial Navigation