NFFLr - Neural Force Field Learning toolkit

Table of Contents

• Introduction
• Installation
• Examples
• How to contribute
• Correspondence
• Funding support

NFFLr (Introduction)

The Neural Force Field Learning library (docs) is intended to be a flexible toolkit for developing and deploying atomistic machine learning systems, with a particular focus on crystalline material property and energy models.

The initial codebase is a fork of ALIGNN, with modified configuration and modeling interfaces for performance.

Installation

We recommend using a per-project pyenv-virtualenv or conda environment.

To ensure proper CUDA support, make sure to install the GPU versions of PyTorch and DGL. For example, to set up a conda environment on linux with with python 3.10 and CUDA 12.1:

conda create --name myproject python=3.10
conda activate myproject
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
conda install -c dglteam/label/cu121 dgl
python -m pip install nfflr

Examples

Under construction here.

How to contribute

We gladly accept pull requests.

For detailed instructions, please see Contributing.md

Correspondence

Please report bugs as Github issues (https://github.com/usnistgov/nfflr/issues) or email to brian.decost@nist.gov.

Funding support

NIST-MGI (https://www.nist.gov/mgi).

Code of conduct

Please see Code of conduct