chem-gp/fande

Fitting molecular Energies, Forces and Chemical shifts and Chemical Discovery with Scalable Gaussian Processes by Stochastic Variational Inference

Description

Fitting molecular Energies, Forces and Chemical shifts with scalable Gaussian Processes trained by stochastic variational inference on multiple GPUs.

Project scope is for those doing research in probabilistic machine learning applied to chemical physics problems

Currently fande uses GPyTorch/PyTorch/Pyro for ML modeling and librascal library for computing invariants.

Examples

...

Currently implemented methods

Install

Install by pip:

pip install git+https://github.com/chem-gp/fande
pip install git+https://github.com/chem-gp/fande --force-reinstall