This repo contains code for optimizing molecules using Bayesian optimization and quantum chemistry computations.
The aim is to perform accelerated local search using an active learning approach. That is, the optimization will guess final bond lengths, angles, and dihedrals based on previous optimizations.
At the moment, the project is experimental.
Build the environment using anaconda:
mamba env create --file environment.yml --forcerun.py provides a simple interface to the code. To optimize cysteine with default arguments. For now, the code expects molecules in XYZ format.
python run.py test/molecules/peroxide.xyz