mattersim

graph-pes allows you fine-tune and use the mattersim series of models in the same way as any other GraphPESModel, either via the Python API:

from graph_pes.interfaces import mattersim
model = mattersim("mattersim-v1.0.0-1m")
model.predict_energy(graph)

… or within a graph-pes-train configuration file:

model:
   +mattersim:
      load_path: "mattersim-v1.0.0-5m"

If you use any mattersim models in your work, please visit the mattersim repository and cite the following:

@article{yang2024mattersim,
    title={MatterSim: A Deep Learning Atomistic Model Across Elements, Temperatures and Pressures},
    author={Han Yang and Chenxi Hu and Yichi Zhou and Xixian Liu and Yu Shi and Jielan Li and Guanzhi Li and Zekun Chen and Shuizhou Chen and Claudio Zeni and Matthew Horton and Robert Pinsler and Andrew Fowler and Daniel Zügner and Tian Xie and Jake Smith and Lixin Sun and Qian Wang and Lingyu Kong and Chang Liu and Hongxia Hao and Ziheng Lu},
    year={2024},
    eprint={2405.04967},
    archivePrefix={arXiv},
    primaryClass={cond-mat.mtrl-sci},
    url={https://arxiv.org/abs/2405.04967},
    journal={arXiv preprint arXiv:2405.04967}
}

Installation

To install graph-pes with support for mattersim models, you need to install the mattersim package. We recommend doing this in a new environment:

conda create -n graph-pes-mattersim python=3.9
conda activate graph-pes-mattersim
pip install mattersim graph-pes

Interface

graph_pes.interfaces.mattersim(load_path='mattersim-v1.0.0-1m')[source]

Load a mattersim model from a checkpoint file, and convert it to a GraphPESModel on the CPU.

Parameters:

load_path (str) – The path to the mattersim checkpoint file. Expected to be one of mattersim-v1.0.0-1m or mattersim-v1.0.0-5m currently.

Return type:

GraphPESModel