Anders Johansson

Hm, this is strange. I'll have a look later today. It could be that if you use a lot of MPI ranks, some of them end up without atoms. This...

You should be able to just copy over the new versions of `pair_nequip.*` to `lammps/src` and recompile.

This should be possible, but you need to do everything within CMake, since we use it to find Torch. See the documentation, but you should essentially add something like `-DPKG_PYTHON=ON...

Hi, Thank you for your interest in Allegro! You may be the first person to try NEB with Allegro, and I have never done NEB in LAMMPS myself, so I...

Hi, Is that all the output, does it not print a stack trace of any kind? As a sanity check to verify that this issue really is Allegro-related: Does your...

Hi, Are you saying that you get this error message with an unmodified `in.neb.sivac` that does not reference Allegro at all? If so, this sounds like a problem with your...

Hi @Huni-ML , The Allegro model simply does not output stress, so this won't work. Alby is working on implementing it. Unlike NequIP, the Allegro model does not get a...

@Huni-ML We usually train on stress and find that it helps with getting the correct lattice parameter. You may, however, need to re-check the DFT convergence, as stress can be...

Hi, This is certainly not the intended behavior and a bug that we'll fix. I think it should work already if using Kokkos. You may, however, also want to avoid...

Hi, For running with LAMMPS, PyTorch should not interact with or need to know anything about MPI, and PyTorch can safely be built with `-DUSE_DISTRIBUTED=OFF`. If your simulation is hanging,...