Search results

No module named 'graph_longrange' with MacePolar foundation model

Status: Open.

#1393 In ACEsuit/mace;

· bismuthadams1 opened on Mar 2, 2026

Hardcoded force loss calculation causes crashes and computational inefficiency

Status: Open.

#1385 In ACEsuit/mace;

· matteotolloso opened on Feb 27, 2026

Early stopping fails to terminate training in non-distributed mode

Status: Open.

#1384 In ACEsuit/mace;

· matteotolloso opened on Feb 27, 2026

kokkos cuda issue when using MACE in LAMMPS with ML-IAP

Status: Open.

#1367 In ACEsuit/mace;

· wudiazhu opened on Feb 12, 2026

Model names should be added to README and <code>mace_mp</code> should have a better alias

Status: Open.

#1365 In ACEsuit/mace;

· Andrew-S-Rosen opened on Feb 11, 2026

Add a <code>torch-dftd</code> as an "extra"

Status: Open.

#1364 In ACEsuit/mace;

· Andrew-S-Rosen opened on Feb 11, 2026

Cleaner handling of partially downloaded checkpoints

Status: Open.

#1363 In ACEsuit/mace;

· Andrew-S-Rosen opened on Feb 11, 2026

NaN temperature and inf pair energy with mliap potential in LAMMPS

Status: Open.

#1353 In ACEsuit/mace;

· park2971 opened on Jan 30, 2026

[BUG] LAMMPS ML-IAP build fails with Python 3.12: Cython-generated mliap_model_python_couple.cpp calls va_start with 1 arg

lammps

Status: Open.

#1349 In ACEsuit/mace;

· evenfarther opened on Jan 28, 2026

On multiply GPUs, the code fails in evaluation after training

Status: Open.

#1348 In ACEsuit/mace;

· gitlzzz opened on Jan 27, 2026

Reactivity in condensed phase: putting higher weights on solutes' atoms compared to solvents' atoms in the training loss.

Status: Open.

#1339 In ACEsuit/mace;

· ines-art opened on Jan 25, 2026

Memory problem when trying to train with large system.

Status: Open.

#1337 In ACEsuit/mace;

· jungsdao opened on Jan 23, 2026