Welcome to mlcg’s documentation!

This is the base code to create models such as the ones described in [TransCGSchnet] and [CGSchnet].

This repository collects a set of tools to apply machine learning techniques to coarse grain atomic systems.

Installation

First we suggest to create a new clean empty virtual environment with python 3.12, then clone the repo and install the following prerequisites:

git clone git@github.com:ClementiGroup/mlcg.git
cd mlcg
pip install -r env_with_hashes.in
pip install --no-deps git+https://github.com/ACEsuit/mace.git@v0.3.13
pip install --no-deps nequip==0.12.1 nequip-allegro==0.7.0

Then install this repository with:

pip install .

This will likely rise an error due to some dependency issue about e3nn that you can safely ignore.

CLI

The models defined in this library can be conveniently trained using the pytorch-lightning CLI utilities.

Examples

Please take a look into the examples folder of the repository to see how to use this code to train a model over an existing dataset.

Contents

• Data and datasets
  • Molecular Data Representation
  • Fixed datasets
  • H5 Dataset
• Geometry and Neighbor lists
  • Internal Coordinate Functions
  • Neighbor Lists
  • Topology Utilities
  • Statistics Utilities
• Model utilities
  • Priors
  • Cutoff Functions
  • Radial Basis Functions
  • MLIPs
  • Gradient/sum wrappers
  • Loss Functions
• Training
• Simulations
  • Langevin Simulation
  • Overdamped Simulation
  • Parallel Tempering Simulation
  • Simulation utilities
• Miscellaneous utilities
  • Coarse Graining
  • General utilities
  • Molecular utilities
• Developer Documentation
  • Documentation
  • Tests
• Examples
• Bibliography

Indices and tables

• Index

• Module Index

• Search Page