nvalchemi Model Wrapping

Overview

To use an arbitrary MLIP (Machine Learning Interatomic Potential) within nvalchemi, wrap it using the BaseModelMixin interface. This standardizes how models receive AtomicData/Batch inputs and produce ModelOutputs.

from nvalchemi.models.base import BaseModelMixin, ModelCard, ModelConfig
from nvalchemi.data import AtomicData, Batch

Architecture

A wrapped model uses multiple inheritance: your PyTorch model class + BaseModelMixin.