IMPACT-ML¶

Intelligent Mass Parametrization for Advanced Classification Tasks with Machine Learning

Setup¶

First, the repository is clone via

git clone --recurse-submodules https://github.com/nshadskiy/IMPACT-ML.git

and after going to the folder (cd IMPACT-ML) the environment is set up via conda

conda env create -f environment.yaml

don't forget to change the conda environment path in the environment.yaml file to your conda installation which is defined as

prefix: /work/USER/CONDAPATH/envs/nn_env

To run the preselection step, execute the python script and specify the config file:

python preselection.py --config-file configs/PATH/CONFIG.yaml

To train a network run:

python trochscript.py -i /ceph/USER/SOME/PATH/FOLDER -c configs/FILE.yaml -y 2018 -k mt -s even -o SOME_OUTPUT_NAME -e 2000