OMNI-P1

OMNI-P1 is the all-in-one universal potential that was simultaneously trained on two levels of theory, DFT (wB97X/def2-TZVPP) and coupled cluster (CCSD(T)*/CBS), see:

Details are in the paper:

• Yuxinxin Chen, Pavlo O. Dral. One to Rule Them All: A Universal Interatomic Potential Learning across Quantum Chemical Levels. J. Chem. Theory Comput., 2025, 21, 8762–8772. DOI: 10.1021/acs.jctc.5c00858. Preprint on ChemRxiv: https://doi.org/10.26434/chemrxiv-2024-ng3ws (2024.09.18) | arXiv: https://arxiv.org/abs/2409.12015.

Note

OMNI-P1 can be fine-tuned on your own data. A few dozen reference points are often enough to specialize it to your system, and the fine-tuned model is used exactly like the original one. See the fine-tuning tutorial.

The procedure is general and can be applied for simultaneous training on more levels of theory.

Due to the uniqueness of this model, you can choose the target level of theory to make predictions at. Default is CCSD(T)*/CBS as a better level. After you choose the level, you can use the model for the typical simulations such as geometry optimizations or molecular dynamics.

Below is a simple example for geometry optimizations:

initmol = ml.molecule.from_xyz_file('init.xyz')
omnip1 = ml.methods(method='omnip1', level='cc')
geomopt = ml.optimize_geometry(model=omnip1, initial_molecule=initmol)
final_mol = geomopt.optimized_molecule
print('Optimized coordinates:')
print(final_mol.get_xyz_string())
final_mol.write_file_with_xyz_coordinates(filename='final.xyz')