instructions.md

python environment install with conda

# create new environment from file
conda create -n "myenv" --file requirements.txt python>=3.9.0
# install sciris
pip install sciris
# install pytorch
conda install pytorch==1.13.1 pytorch-cuda=11.6 -c pytorch -c nvidia

Information of each folder:

/covasim contains the source code for the agent-based simulation model Covasim
/Notebooks contains the source code for BINN
/tests contains the source code for data generation

Sequence of execution for BINN

BINNCovasimTraining.py for model training

# after model training, remember to replace my_dir with the path for the trained model

BINNCovasimParameterNN_dynamic.py for learning symbolic functions for each parameter neural network

BINNCovasimEvaluation_dynamic.py for evaluation and comparison for ODE models