Learn TensorFlow, use the source code to install _tensorflow

PC installed Ubuntu System, we step-by-step to explain how to use the source code to install TensorFlow? (my Ubuntu system is 15.10) install Cuda

According to your system model select the appropriate cuda version of the download

Https://developer.nvidia.com/cuda-downloads

I chose Cuda-repo-ubuntu1504-7-5-local_7.5-18_amd64.deb.

Right-click the downloaded file, Open interminal, and enter the command:

sudo dpkg-icuda-repo-ubuntu1504-7-5-local_7.5-18_amd64.deb

sudo apt-get update

sudo apt-get install Cuda

Install to:/usr/local/cuda

Download CUDNN,HTTPS://DEVELOPER.NVIDIA.COM/CUDNN

My next version is cudnn-7.0-linux-x64-v4.0-prod.tgz.

Ditto, enter the following command:

Tar xvzf cudnn-7.0-linux-x64-v4.0-prod.tgz

sudo cp cuda/include/cudnn.h/usr/local/cuda/include

sudo cp cuda/lib64/libcudnn*/usr/local/cuda/lib64

sudo chmod a+r/usr/local/cuda/lib64/libcudnn*

Install TensorFlow

Download TensorFlow source files

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

Installation Bazel,bazel is the Google Open Source build tool, similar to make tool, used to compile the build TensorFlow

Http://bazel.io/docs/install.html

Bazel Download Address:

Https://github.com/bazelbuild/bazel/releases (bazel-0.2.2b-installer-linux-x86_64.sh)

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

chmod +x bazel-version-installer-os.sh

./bazel-version-installer-os.sh--user

Export path= "$PATH: $HOME/bin"

Installation TensorFlow Other dependencies

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

Configure TensorFlow, you need to specify the appropriate file installation directory. CD into TensorFlow source file.

sudo./configure

When configured, the following will appear:

Setting up Cuda include

Setting up Cuda lib64

Setting up Cuda Bin

Setting up Cuda NVVM

Setting up Cupti include

Setting up Cupti lib64

Configuration finished

Building with Bazel compilation

Bazel build-c opt--config=cuda//tensorflow/cc:tutorials_example_trainer

Bazel-bin/tensorflow/cc/tutorials_example_trainer--use_gpu

# Lots of output. This is tutorial iteratively calculates the major eigenvalue
of # a 2x2 matrix, on GPU. The last few lines is like this.
000009/000005 lambda = 2.000000 x = [0.894427-0.447214] y = [1.788854-0.894427]
000006/000001 lambda = 2.000000 x = [0.894427-0.447214] y = [1.788854-0.894427]
000009/000009 lambda = 2.000000 x = [0.894427-0.447214] y = [1.788854- 0.894427]

Ok!

Test:

Cd/home/lei/tensorflow/tensorflow/models/image/mnist

Python convolutional.py

Tips:

Configure, GCC version <=4.9 CUDNN Version = V4

Resources:

Installing from sources one problem