search

MDIL-SNU / SevenNet

SevenNet - a graph neural network interatomic potential package supporting efficient multi-GPU parallel molecular dynamics simulations.
https://pubs.acs.org/doi/10.1021/acs.jctc.4c00190
GNU General Public License v3.0
105 stars 13 forks source link

Publish to PyPI #52

Closed janosh closed 1 month ago

janosh commented 1 month ago

opening this to keep track of the PyPI release for SevenNet mentioned here

YutackPark commented 1 month ago

Thank you for your patience. SevenNet is now released on PyPI. I have replaced torch_scatter, which required manual installation steps, with pure PyTorch functions. With PyTorch installed, SevenNet can now be installed via pip.

You may be interested in pre-trained models. Our model, which is around 10 MB, is included in the project itself as a data file. It can be loaded, for example, as an ASE calculator:

from sevenn.sevennet_calculator import SevenNetCalculator
sevenet_0_cal = SevenNetCalculator("7net-0_11July2024", device='cpu')

Let me know if you need any further suggestions!