Contents Menu Expand Light mode Dark mode Auto light/dark, in light mode Auto light/dark, in dark mode Skip to content
KLIFF 1.0.1 documentation
Logo

The Basics

  • Installation
  • Introduction
    • Practical Introduction to the Dataset Module
    • Weights
    • Trainer Manifest
    • Example: Training a KIM Potential
    • Transforms
    • Example: Training a Descriptor based Potential
    • Example: Training a Graph neural netwok based Potential
  • Theory
    • Theory
    • Graph Convolutions in KLIFF

Advanced Topics

  • Running the model in LAMMPS or ASE
  • How To
    • Save and load a model
    • Install a model
    • Implement a new model
    • Run in parallel mode
  • Command Line Tool
  • Contributing guide

UQ and Legacy Modules

  • Legacy Module
  • Frequently Used Modules
    • Dataset
    • Models
    • Model Class Documentation
    • KIM models
    • Neural network models
    • Descriptors
    • Calculators
    • Loss
    • Uncertainty Quantification (UQ)
  • Tutorials
    • Train a Stillinger-Weber potential
    • Train a neural network potential
    • Train a neural network potential for SiC
    • Parameter transformation for the Stillinger-Weber potential
    • MCMC sampling
    • Bootstrapping
    • Train a Lennard-Jones potential
    • Train a linear regression potential

Extra Information

  • Change Log
  • Frequently Asked Questions
  • Package Reference
    • kliff.analyzers
    • kliff.analyzers.fisher
    • kliff.analyzers.rmse
    • kliff.atomic_data
    • kliff.dataset
    • kliff.dataset.configuration
    • kliff.dataset.dataset
    • kliff.dataset.dataset_torch
    • kliff.dataset.extxyz
    • kliff.dataset.weight
    • kliff.error
    • kliff.legacy
    • kliff.legacy.calculators
    • kliff.legacy.calculators.calculator
    • kliff.legacy.calculators.calculator_torch
    • kliff.legacy.descriptors
    • kliff.legacy.descriptors.bispectrum
    • kliff.legacy.descriptors.bispectrum.bispectrum
    • kliff.legacy.descriptors.descriptor
    • kliff.legacy.descriptors.symmetry_function
    • kliff.legacy.descriptors.symmetry_function.sym_fn
    • kliff.legacy.loss
    • kliff.legacy.nn
    • kliff.log
    • kliff.models
    • kliff.models.kim
    • kliff.models.lennard_jones
    • kliff.models.linear_regression
    • kliff.models.model
    • kliff.models.model_torch
    • kliff.models.neural_network
    • kliff.models.parameter
    • kliff.neighbor
    • kliff.neighbor.neighbor
    • kliff.parallel
    • kliff.trainer
    • kliff.trainer.base_trainer
    • kliff.trainer.kim_trainer
    • kliff.trainer.lightning_trainer
    • kliff.trainer.torch_trainer
    • kliff.trainer.utils
    • kliff.trainer.utils.dataloaders
    • kliff.trainer.utils.lightning_utils
    • kliff.trainer.utils.losses
    • kliff.transforms
    • kliff.transforms.configuration_transforms
    • kliff.transforms.configuration_transforms.configuration_transform
    • kliff.transforms.configuration_transforms.default_hyperparams
    • kliff.transforms.configuration_transforms.descriptor_initializers
    • kliff.transforms.configuration_transforms.descriptors
    • kliff.transforms.configuration_transforms.graphs
    • kliff.transforms.configuration_transforms.graphs.generate_graph
    • kliff.transforms.configuration_transforms.utils
    • kliff.transforms.parameter_transforms
    • kliff.transforms.property_transforms
    • kliff.uq
    • kliff.uq.bootstrap
    • kliff.uq.mcmc
    • kliff.uq.mcmc_utils
    • kliff.utils
  • GitHub Repository
Back to top
View this page

TheoryΒΆ

Fundamental principles for training interatomic potentials,

  • Theory
  • Graph Convolutions in KLIFF
Next
Theory
Previous
Example: Training a Graph neural netwok based Potential
Copyright © 2021-2023, OpenKIM
Made with Sphinx and @pradyunsg's Furo