Welcome to KLay’s documentation!

Contents:

• KLay – composable layers for MLIPs
  • Why another library?
  • Key features
  • Quick installation
  • One-minute example
  • Architecture in a nutshell
  • Command-line highlights
• Using KLay
  • 1. model_params – hyper-parameters & shared constants
  • 2. model_inputs – declare the forward() signature
  • 3. model_layers – the heart of the graph
  • 4. model_outputs – what you want out
  • Cheat-sheet
  • Gotchas & best practices
• Training a MACE-style GNN with klay + kliff
  • Prerequisites
  • Next steps
• Advanced Models
  • End-to-end example
  • 1. Arbitrary Layers (ArbitraryModule)
  • 2. Layer Aliases (weight sharing & staged graphs)
  • Caveats & compliance notes

Indices and tables

• Index

• Module Index

• Search Page