_images/upet-logo-with-text.svg _images/upet-logo-with-text-dark.svg

UPET is a family of universal interatomic potentials for advanced materials modeling across the periodic table. These models are based on the Point Edge Transformer (PET) architecture, trained on a variety of popular atomistic datasets, and capable of predicting energies and forces in complex atomistic workflows.

The package also ships PET-MAD-DOS, a universal model for predicting the electronic density of states (DOS) of materials and molecules, as well as their Fermi levels and bandgaps. PET-MAD-DOS uses a slightly modified PET architecture and is trained on the MAD dataset.

Note

The PET-MAD-1.6 models, trained for 102 elements at the r2SCAN level of theory, are now available. On top of the MAD-1.5 training data, they were additionally trained on catalytic surfaces, and therefore have better accuracy for surface reactions and adsorption energies. They also come in a new M size. See Available models and the updated preprint for more details.

from upet.ase import UPETCalculator
calculator = UPETCalculator(model="pet-mad-s", version="1.6.0", device="cuda")

Note

A new experimental integration of UPET with the NVIDIA ALCHEMI Toolkit is now available. It allows for GPU-native batched inference, relaxations and MD simulations with a torch.compiled version of UPET. See NVIDIA ALCHEMI Toolkit.

from upet.nvalchemi import UPETWrapper
model = UPETWrapper.from_checkpoint(
    model="pet-mad-s", version="1.6.0", device="cuda"
)

Key features

  • Universality: UPET models are generally applicable, and can be used for predicting energies and forces, as well as the density of states, Fermi levels, and bandgaps for a wide range of materials and molecules.

  • Accuracy: UPET models achieve excellent accuracies in various types of atomistic simulations of organic and inorganic systems.

  • Efficiency: UPET models are highly computationally efficient and have low memory usage, which makes them suitable for large-scale simulations.

  • Infrastructure: Various MD engines are available for diverse research and application needs.

  • HPC compatibility: Efficient in HPC environments for extensive simulations.

Maintainers

This project is maintained by @abmazitov, who will reply to issues and pull requests opened on the repository as soon as possible. You can mention them directly if you have not received an answer after a couple of days.