Sunday, September 11, 2016

Install notes for Tensorflow

Installed CUDA 7.5 and CuDNN 5.1.3

Steps modified from

# Need Java 1.8 for bazel, on 14.04 this is a big mess - 14.10 and up openjdk-8 is in the repos
sudo dpkg --install --recursive --auto-deconfigure libjpeg62-turbo_1.3.1-12_amd64.deb
sudo apt-get update

sudo apt-get install debian-keyring debian-archive-keyring

sudo sh -c 'echo "deb jessie-backports main" >> /etc/apt/sources.list.d/debian-jessie-backports.list'

sudo sh -c 'echo "Package: *" >> /etc/apt/preferences.d/debian-jessie-backports'

sudo sh -c 'echo "Pin: release o=Debian,a=jessie-backports" >> /etc/apt/preferences.d/debian-jessie-backports'

sudo sh -c 'echo "Pin-Priority: -200" >> /etc/apt/preferences.d/debian-jessie-backports'
sudo apt-get -t jessie-backports install openjdk-8-jre openjdk-8-jre-headless openjdk-8-jdk-headless openjdk-8-jdk

# Set both to Java 8
sudo update-alternatives --config java
sudo update-alternatives --config javac

# Now for bazel
git clone
cd bazel
git checkout tags/0.2.1
sudo cp output/bazel /usr/bin

# Tensorflow
git clone --recurse-submodules
git checkout r0.10
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64"
export CUDA_HOME=/usr/local/cuda
bazel build -c opt --config=cuda //tensorflow/tools/pip_package:build_pip_package
pip install wheel
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
pip install /tmp/tensorflow_pkg/tensorflow-0.10.0-cp27-none-linux_x86_64.whl

Saturday, September 10, 2016

Get CuDNN from command line

Check the Dockerfile

Monday, August 1, 2016

Magic fix for GPU issues in Tensorflow

E tensorflow/stream_executor/cuda/] failed call to cuInit: CUDA_ERROR_UNKNOWN
I tensorflow/stream_executor/cuda/] libcuda reported version is: Not found: was unable to find DSO loaded into this program

I tensorflow/stream_executor/cuda/] driver version file contents: """NVRM version: NVIDIA UNIX x86_64 Kernel Module  367.35  Mon Jul 11 23:14:21 PDT 2016
GCC version:  gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.3)
I tensorflow/stream_executor/cuda/] kernel reported version is: 367.35.0
I tensorflow/core/common_runtime/gpu/] No GPU devices available on machine.

This error can be fixed by

sudo apt-get install nvidia-modprobe

See issue

Friday, June 17, 2016

Screen bindings to start paned session and rebind pane focus

Put the following lines in ~/.screenrc

bind j focus down                                                            
bind k focus up                                                              
bind l focus right                                                          
bind h focus left                                                            
layout new                                                                    
split -v                                                                      
screen 0                                                                      
focus right                                                                  
screen 1                                                                      
focus down

Wednesday, May 18, 2016

SOCKS proxy with chrome and SSH

In the particular case of an ssh proxy running, connected to something like ssh -D 4321

google-chrome --proxy-server="socks5://" --host-resolver-rules="MAP * , EXCLUDE"


 /Applications/Google\\ Chrome --proxy-server="socks5://" --host-resolver-rules="MAP * , EXCLUDE"

Friday, March 25, 2016

Skype on Ubuntu 14.04

sudo dpkg --add-architecture i386
sudo sh -c "echo 'deb trusty partner' >> /etc/apt/sources.list.d/canonical_partner.list"
sudo apt-get update
sudo apt-get install skype
LD_PRELOAD=/usr/lib/i386-linux-gnu/mesa/ skype

Friday, March 18, 2016

Install xvfb Ubuntu

sudo apt-get install xvfb x11-xkb-utils xfonts-100dpi xfonts-75dpi xfonts-scalable xfonts-cyrillic x11-apps

xvfb-run app

 xvfb-run -a python

 xvfb-run -a -e /dev/stdout python -u 2>&1