Skip to main content
Ctrl+K

chemtrain

  • Documentation Overview

Algorithms

  • Differentiable Trajectory Reweighting (DiffTRe)
  • Force Matching
  • Parallel force matching
  • Distributed DiffTRe
  • Relative Entropy Minimization
  • Prior Simulation

Examples

  • DiffTRe: CG Water on Structural Data
  • Alanine Dipeptide in Implicit Water
  • Bottom-Up and Top-Down Training of Atomistic Titanium

chemtrain

  • compose
    • compose.mace_jax
    • compose.utils
  • data
    • data.preprocessing
    • data.data_loaders
  • deploy
    • deploy.exporter
    • deploy.comm
    • deploy.graphs
  • learn
    • learn.difftre
    • learn.force_matching
    • learn.probabilistic
    • quantity.property_prediction
    • learn.max_likelihood
  • quantity
    • quantity.property_prediction
    • quantity.targets
    • quantity.observables
    • quantity.constants
  • ensemble
    • ensemble.evaluation
    • ensemble.reweighting
    • ensemble.sampling
      • ensemble.sampling.TimingClass
      • ensemble.sampling.SimulatorState
      • ensemble.sampling.TrajectoryState
  • trainers
  • typing

chemtrain-deploy

  • Installation
  • Getting Started
  • LAMMPS Package
  • Model Inputs
  • MACE Foundation Models

Adapter Development

  • JCN Adapter API

Jax, M.D. Extensions

  • model
    • model.prior
    • model.neural_networks
    • model.dimnet_basis_util
    • model.layers
    • model.sparse_graph
  • custom_electrostatics
  • custom_energy
    • jax_md_mod.custom_energy.lorentz_berthelot
  • custom_interpolate
  • custom_partition
  • custom_quantity
  • custom_space
  • io
  • .rst

compose

compose#

  • compose.mace_jax
  • compose.utils

previous

Bottom-Up and Top-Down Training of Atomistic Titanium

next

compose.mace_jax

By Multiscale Modeling of Fluid Materials

© Copyright 2026, Multiscale Modeling of Fluid Materials, TU Munich.