ML_Calc_Driver is a driver to emulate DFT calculations using machine learned predictive models. To use this package, one needs an already trained model, that can predict energy or forces from an input atomic geometry. Supported systems depends on the model, this package should not be a limitation.
This package has been tested with python 3.7, but should work with older python 3 versions.
Credit to mmoriniere for the MyBigDFT package which served as a foundation for this work. Some classes have been directly adapted from MyBigDFT.
Dependencies scanned by PyUp.io
The documentation is available here.
To install the latest version available on PyPi:
pip install mlcalcdriver
To upgrade to last version:
pip install -U mlcalcdriver
git clone git@github.com:OMalenfantThuot/ML_Calc_Driver.git
cd ML_Calc_Driver
pip install .
To modify the sources, instead of pip install ., use
pip install -r requirements_dev.txt
pip install -e .
To use with SchetPack trained models (only supported models at the time)
pip install schnetpack