MLPotential Manual 2021.1

MLPotential is an interface between the Amsterdam Modeling Suite and several types of machine learning potential backends.

• PDF version of the MLPotential Manual

Table of Contents:

• Theory and usage
  • Quickstart guide
  • Theory of ML potentials
  • Installation and uninstallation
  • Included (pre-parameterized) models
  • Custom models (custom parameters)
  • Backends
  • CPU and GPU (CUDA), parallelization
  • Troubleshooting
  • Support
  • Technical information
  • References
• AMS driver’s tasks and properties
  • Geometry, System definition
  • Tasks: exploring the PES
  • Properties in the AMS driver
• MLPotential Keywords

• Index