This subfolder contains the scripts needed to reproduce the argon results.
An installation of LAMMPS is required (e.g., sudo apt install lammps).
The dataset can be generated by running the "launch.sh" script. Then, processing
can be carried out with the "process.py" script. Finally, the "train.sh" script can
be used to train the model.

A technical note: LAMMPS saves the particles of argon as having type 1. ASE reads
these as hydrogens, resulting in no mass-scaling of positions and momenta. As a result,
the errors in the training log file correspond to the raw (non-mass-scaled) errors on
positions and momenta, which can be compared with the MDNet results from the literature.
