INSTALL.md

**If you are running experiments on GPU clusters, do not forget to check dependencies of the installed libraries with available cuda and gcc modules

activate modules

• module load cuda/8.0.44
• module load cudnn/8.0-v5.1
• module load gcc/5.2.0
• module load git

get and install conda

• wget https://repo.continuum.io/miniconda/Miniconda2-latest-Linux-x86_64.sh
• chmod +x Miniconda2-latest-Linux-x86_64.sh
• ./Miniconda2-latest-Linux-x86_64.sh

logout - login to activate conda (or source .bashrc)

• exit

create and active conda virtual environment

• conda create -n tfenv
• source activate tfenv

Add conda-forge repository, install tensorflow library

• conda config --add channels conda-forge
• conda install tensorflow-gpu

install required libraries

• conda install ipython
• conda install scipy
• conda install scikit-learn
• conda install h5py
• pip install wget
• conda install matplotlib
• pip install -U nltk
• python -m nltk.downloader -d /(your-home-directory)/nltk_data all
• conda install gensim

** install seaborn (because I love their visualization)

** this will also instantly install python pandas - or can also be installed prior with conda install pandas

• cd (your-git-directory)
• git clone https://github.com/mwaskom/seaborn.git
• cd seaborn
• python setup.py install

download and install keras

• cd (your-git-directory)
• git clone https://github.com/fchollet/keras.git
• cd keras
• python setup.py install pip install git+git://github.com/fchollet/keras.git --upgrade

Test Installation

python

import tensorflow as tf

graph = tf.constant('Hello world')
session = tf.Session()
print(session.run(graph))
session.close()

deactivate/quit conda tensorflow environment

source deactivate tfenv