Model and Data from: A Lightweight Universal Machine-Learning Interatomic Potential via Knowledge Distillation for Scalable Atomistic Simulations

This repository contains the dataset, models, and code supporting the paper: “A Lightweight Universal Machine-Learning Interatomic Potential via Knowledge Distillation for Scalable Atomistic Simulations”.

Contents included:

• checkpoints.tar: Pre-trained checkpoints of the 7net-Nano model.

• Modified SevenNet packages for fine-tuning.

• example.tar: Example code for fine-tuning applied to liquid electrolyte applications.

• dft.tar: DFT calculation data used for benchmarking SiO2 with CFx plasma etching simulations