Universal Potential: M3GNet-UP-2022¶
Note: This example requires AMS2023 or later.
This example illustrates how set up an AMS calculation using the M3GNet 2022 Universal Potential from the MLPotential Engine.
Example usage: (Download M3GNet.py
)
#!/usr/bin/env amspython
from scm.plams import *
def main():
mol = from_smiles("O")
mol.lattice = [
[
3.0,
0.0,
0.0,
],
[0.0, 3.0, 0.0],
[0.0, 0.0, 3.0],
]
s = Settings()
s.runscript.nproc = 1
s.input.ams.task = "GeometryOptimization"
s.input.ams.GeometryOptimization.Convergence.Gradients = 0.01 # hartree/ang
s.input.MLPotential.Model = "M3GNet-UP-2022"
# If you have trained a custom M3GNet model yourself, you can use:
# s.input.MLPotential.Model = 'Custom'
# s.input.MLPotential.Backend = 'm3gnet'
# s.input.MLPotential.ParameterDir = '<path to directory containing the M3GNet model>'
job = AMSJob(settings=s, molecule=mol, name="ams_with_m3gnet")
job.run()
energy = job.results.get_energy(unit="eV")
print(f"M3GNet: final energy {energy:.3f} eV")
if __name__ == "__main__":
main()