Machine Learning Software Setup

Python

Install Python 2.7 from scratch, so that we don't have to worry about lxplus-speific python versions/configuration/environment. This takes some time, but gives us a stable environment to work in. It will be located in ~/.localpython.

export MY_ML_DIR=$HOME/ML
mkdir -p $MY_ML_DIR/python
cd $MY_ML_DIR/python
wget http://www.python.org/ftp/python/2.7.9/Python-2.7.9.tgz
tar -zxvf Python-2.7.9.tgz
cd Python-2.7.9
mkdir ~/.localpython
./configure --prefix=$HOME/.localpython
make
make install

Make virtual environment so that we work in a clean, self contained python setup:

cd $MY_ML_DIR/python
wget https://pypi.python.org/packages/5c/79/5dae7494b9f5ed061cff9a8ab8d6e1f02db352f3facf907d9eb614fb80e9/virtualenv-15.0.2.tar.gz#md5=0ed59863994daf1292827ffdbba80a63
tar -zxvf virtualenv-15.0.2.tar.gz
cd virtualenv-15.0.2/
~/.localpython/bin/python setup.py install
virtualenv ve -p $HOME/.localpython/bin/python2.7
source ve/bin/activate   

# Initialize virtual Python environment
virtualenv py_virtualenv

# This activates the environment! Has to be run after each login
source py_virtualenv/bin/activate

# ML and ML-related libraries (you may not need all of these)
pip install theano
pip install keras=="1.2.0"
pip install h5py
pip install pypng
pip install sklearn

{
    "epsilon": 1e-07,
    "floatx": "float32",
    "image_data_format": "channels_last",
    "backend": "theano"
}

Resources

https://indico.cern.ch/event/595059/contributions/2522193/attachments/1430921/2197986/slides_iml_keras_workshop.pdf

https://root.cern.ch/doc/v608/dir_050d74ad7a64b0aa8b3fbebce21e7998.html