modelforge

This package is centered around the implementation and infrastructure to train, create, optimize and store neural network potentials (NNPs) effectively. Datasets are provided to enable accurate training and validation of the neural network structures, and users can include their own using the API in modelforge-curate. The technical roadmap for the modelforge package is outlined in the wiki.

Documentation for how to use the package can be found at: https://modelforge.readthedocs.io/en/latest/

Subpackages

• modelforge: Core infrastructure for training, evaluating, and managing neural network potentials.
• modelforge-curate: Dataset curation API and tools for building standardized datasets.
• modelforge-ase: ASE calculator wrapper for running modelforge potentials through ASE workflows.
• modelforge-openmm: OpenMM wrapper for using modelforge models in OpenMM simulations.

Installation

To set up any environment, first clone the repository:

git clone https://github.com/choderalab/modelforge.git

navigate to the modelforge root directory:

cd modelforge

Install Core Package (modelforge)

Create and activate the base environment:

micromamba create -n modelforge -f devtools/conda-envs/env.yaml
micromamba activate modelforge

Install the core package:

pip install -e . --no-deps --config-settings editable-mode=strict

Install Dataset Curation Subpackage (modelforge-curate)

From the base modelforge environment:

pip install -e modelforge-curate --no-deps --config-settings editable-mode=strict

Install ASE Integration Subpackage (modelforge-ase)

Create and activate the ASE runtime environment:

micromamba create -n modelforge-ase -f devtools/conda-envs/env_modelforge_ase.yaml
micromamba activate modelforge-ase

Install modelforge and the ASE integration package:

pip install -e . --no-deps --config-settings editable-mode=strict
pip install -e modelforge-ase --no-deps --config-settings editable-mode=strict

Install ASE Examples/Notebook Dependencies

If you want to run ASE notebooks/examples (includes ase, ipykernel, nglview):

micromamba create -n modelforge-ase-examples -f devtools/conda-envs/env_modelforge_ase_examples.yaml
micromamba activate modelforge-ase-examples
pip install -e . --no-deps --config-settings editable-mode=strict
pip install -e modelforge-ase --no-deps --config-settings editable-mode=strict
python -m ipykernel install --user --name modelforge-ase-examples --display-name "Python (modelforge-ase-examples)"

Install OpenMM Integration Subpackage (modelforge-openmm)

Create and activate the OpenMM runtime environment:

micromamba create -n modelforge-openmm -f devtools/conda-envs/env_modelforge_openmm.yaml
micromamba activate modelforge-openmm

Install modelforge and the OpenMM integration package:

pip install -e . --no-deps --config-settings editable-mode=strict
pip install -e modelforge-openmm --no-deps --config-settings editable-mode=strict

Note: Test environments in devtools/conda-envs/test_env*.yaml are intended for CI/development testing and are intentionally separate from runtime/example environments.

Copyright

Copyright (c) 2023-2026, Chodera Lab https://www.choderalab.org/

Acknowledgements

Project based on the Computational Molecular Science Python Cookiecutter version 1.1.