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

Contents:

  • Installation
  • User guide
    • Dataset preparation
    • Model training
    • Simulations
    • Advanced
      • Data processing
      • Models
      • Advanced model training
  • API reference
    • Data processing
      • Graph Dataset Builder
      • Chemical System Datasets
        • Dataset interface
        • HDF5 Dataset
        • Reader Utilities
      • Chemical System
      • Dataset Info
      • Dataset Configs
      • Atomic Number Table
      • Graph Definition
      • Graph Dataset
      • Data parallelization and batch prefetching
      • Data split
      • Helper functions
    • Models
      • Base class
      • MACE
      • NequIP
      • ViSNet
      • LiTEN
      • EquiformerV2
      • UMA
      • Force Field predictor
      • Prediction
    • Layers
      • Commonly used layers
      • ESCN Layers
    • Model training
      • Training Loop
      • Training step
      • Optimizer
      • IO handling during training
      • Model evaluation
    • Loss
      • Loss
    • Simulations
      • Simulation Engine: Interface
      • Simulation Config
      • Simulation State
      • Simulation Enums
      • Create graph from ASE atoms
      • JAX-MD Simulation Engine
      • ASE Simulation Engine
      • MLIP ASE Calculator
      • Temperature scheduling
      • Batched inference
    • Utils
      • Model I/O
Back to top
View this page

Model trainingΒΆ

This module contains all code related to model training.

  • Training Loop
  • Training step
  • Optimizer
  • IO handling during training
  • Model evaluation
Next
Training Loop
Previous
ESCN Layers
Copyright © 2025, InstaDeep & Cao Bohan
Made with Sphinx and @pradyunsg's Furo