Tutorials#
To learn how to use KLIFF, begin with the tutorials.
Train a Stillinger-Weber potential: a good entry point to see the basics of training a physics-motivated potential.
Parameter transformation for the Stillinger-Weber potential: it is similar to Train a Stillinger-Weber potential, except that some parameters are transformed to the log space for optimization.
Train a neural network potential: walks through the steps to train a machine-learning neural network potential.
Train a neural network potential for SiC: similar to Train a neural network potential, but train for a system of multiple species.
Train a Lennard-Jones potential: similar to Train a Stillinger-Weber potential (where a KIM model is used), here the Lennard-Jones model built in KLIFF is used.
More examples can be found at https://github.com/openkim/kliff/tree/master/examples.
Parameter transformation for the Stillinger-Weber potential
