High-performance training and inference for deep equivariant interatomic potentials

Digital Discovery 2026 Open access

Read full text
Authors: Chuin Wei Tan, Marc Descoteaux, Mit Kotak, Gabriel de Miranda Nascimento, Seán R. Kavanagh, Laura Zichi, Menghang Wang, Aadit Saluja, Yizhong R. Hu, Tess Smidt, Anders Johansson, William C. Witt
+2 more Boris Kozinsky, Albert Musaelian
DOI: 10.1039/d5dd00423c
Cited by: 10
References: 58

Abstract

The NequIP framework is redesigned for scalable distributed training and PyTorch 2.0 compilation. AOT Inductor inference and optimized Allegro kernels accelerate molecular dynamics by factors of 5–18 on practical system sizes.

Fields of Study & Subjects

Machine Learning in Materials Science Quantum many-body systems Model Reduction and Neural Networks

Publication details

Journal / source
Digital Discovery
Publisher
Royal Society of Chemistry (RSC)
Volume
5
Issue
4
Pages
1558–1567
Publication date
1 January 2026
Language
EN