CederGroupHub / chgnet

Pretrained universal neural network potential for charge-informed atomistic modeling https://chgnet.lbl.gov
https://doi.org/10.1038/s42256-023-00716-3
Other
215 stars 55 forks source link

problem with ase.filters #107

Closed Atefeh-Yadegarifard closed 6 months ago

Atefeh-Yadegarifard commented 7 months ago

Email (Optional)

No response

Version

0.3.3

Which OS(es) are you using?

What happened?

Hello. I have installed the latest version of chgnet and I am unable to initialize it. chgnet version 0.3.3 ase version 3.22.1

Code snippet

from chgnet.model.model import CHGNet

Log output

ModuleNotFoundError                       Traceback (most recent call last)
Cell In[1], line 1
----> 1 from chgnet.model.model import CHGNet

File ~/miniconda3/lib/python3.11/site-packages/chgnet/model/__init__.py:3
      1 from __future__ import annotations
----> 3 from chgnet.model.dynamics import CHGNetCalculator, MolecularDynamics, StructOptimizer
      4 from chgnet.model.model import CHGNet
      6 __all__ = ["CHGNet", "StructOptimizer", "MolecularDynamics", "CHGNetCalculator"]

File ~/miniconda3/lib/python3.11/site-packages/chgnet/model/dynamics.py:13
     11 from ase import Atoms, units
     12 from ase.calculators.calculator import Calculator, all_changes, all_properties
---> 13 from ase.filters import Filter, FrechetCellFilter
     14 from ase.md.npt import NPT
     15 from ase.md.nptberendsen import Inhomogeneous_NPTBerendsen, NPTBerendsen

ModuleNotFoundError: No module named 'ase.filters'

Code of Conduct

BowenD-UCB commented 6 months ago

Hi,

The issue comes from the dependency in ase.filters Please refer to ase install in the release Before the official release of ASE pypi package, the installation can only be accessable from gitlab.

The update in the v0.3.3 is to implemente the new FrechetCellFilter in StructOptimizer for better relaxation accuracy and speed.