DeepMETEOR: Deep Learning-based MET Estimation for Online Reconstruction

Recipes

install

You can install python dependencies using conda.

condaa create -y -f environment.yaml

T3_KR_KNU

CONDA_OVERRIDE_CUDA=11.7 mamba create -y -f environment.yaml

setup

setup.zsh exports two environment variables named PROJECT_DATA_DIR and PROJECT_LOG_DIR. PROJECT_DATA_DIR is a directory where datasets are stored and PROJECT_LOG_DIR is a directory where training output files are saved. Edit them to suit your system.

source setup.zsh

preparing datasets

You can download ttbar samples using the following script.

./scripts/download-files.zsh

quick start

All configuration is handled using yaml file.

python train.py ./config/transformer.yaml

slowmoyang / deepmeteor