I am reaching out to seek assistance in utilizing the GAP machine learning potential that developed 2018 publication titled "Achieving DFT accuracy with a machine-learning interatomic potential: Thermomechanics and defects in bcc ferromagnetic iron."
I have taken the initiative to install QUIP into LAMMPS as per the instructions. However, I am encountering difficulties when configuring the pair_style and pair_coeff parameters. Specifically, I have followed the format 'QUIP filename; QUIP initialization string; N atomic numbers = mapping of QUIP elements to atom types.' In my case, I set it as:
pair_style quip
pair_coeff * * gp33b.xml ''IP GAP label=GAP_2016_10_3_60_19_29_10_891'' 1
The reason I configured the QUIP initialization string in this manner is because I noticed in the gp33b.xml file that it mentions init_args="IP GAP label=GAP_2016_10_3_60_19_29_10_891". Additionally, as the fitting is specifically for the potential of iron elements, I set N to 1.
I would greatly appreciate any guidance you can provide to resolve this issue. If there are specific details or steps that I might have overlooked, please point them out. Your assistance is crucial for the progress of my research.
- Badges
- Science topic