Scholarlink

Abstract <p>Development of machine learning tools makes possible to apply them in various fields of science. For example, by creating and training specially designed neural network it is possible to obtain a significant acceleration of computational process compared to ab initio molecular dynamics. This paper presents a Python programming package for creating, training, and applying Behler–Parrinello neural networks using PyTorch framework.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

NNMD: Python Programming Package for Machine-Learning Potentials Training with Use of Behler–Parrinello Functions

  • 期刊论文

  • 发表日期：

    2026年7月20日

查看全文
  • A. S. Budnikov,
  • A. V. Romanov

摘要

Abstract

Development of machine learning tools makes possible to apply them in various fields of science. For example, by creating and training specially designed neural network it is possible to obtain a significant acceleration of computational process compared to ab initio molecular dynamics. This paper presents a Python programming package for creating, training, and applying Behler–Parrinello neural networks using PyTorch framework.