flaviucipcigan/matgfn

Introduction

matgfn is a Python package to train and sample GFlowNets. The gflow module implements Generative Flow Networks using PyTorch and Gymnasium.

The reticular module implements utility functions to generate reticular materials using Secondary Building Units and predict their properties based on features or atom positions and graphs. The notebooks folder shows example notebooks for both use-cases.

Results are reported in Discovery of New Reticular Materials for Carbon Dioxide Capture using GFlowNets.

Citation

@inproceedings{
    cipcigan2023discovery,
    title={Discovery of Novel Reticular Materials for Carbon Dioxide Capture using {GF}lowNets},
    author={Flaviu Cipcigan and Jonathan Booth and Rodrigo Neumann Barros Ferreira and Carine Ribeiro Dos Santos and Mathias B Steiner},
    booktitle={AI for Accelerated Materials Design - NeurIPS 2023 Workshop},
    year={2023},
    url={https://openreview.net/forum?id=cq2MJtq9iA}
}

Installation

# Ensure you have Python>=3.11
pyenv local 3.11

# Clone repository via either SSH or HTTPS
git clone $REPOSITORY_PATH

# Create virtual environment
python -m venv .venv

# Activate virtual environment
source .venv/bin/activate

# Upgrade pip if current pip version is an old version
pip install --upgrade pip

# This pulls all dependencies
pip install -e matgfn