SevenNet

SevenNet is a graph neural network interatomic potential package that supports efficient multi-GPU parallel molecular dynamics simulations. It provides pretrained universal potentials and enables large-scale simulations with excellent pa…

6. DYNAMICS 6.3 Machine Learning Potentials VERIFIED
Back to Directory Official Website

Overview

SevenNet is a graph neural network interatomic potential package that supports efficient multi-GPU parallel molecular dynamics simulations. It provides pretrained universal potentials and enables large-scale simulations with excellent parallel scaling.

Reference Papers

Reference papers are not yet linked for this code.

Full Documentation

Official Resources

  • Homepage: https://github.com/MDIL-SNU/SevenNet
  • Documentation: https://sevennet.readthedocs.io/
  • Source Repository: https://github.com/MDIL-SNU/SevenNet
  • License: GPL-3.0

Overview

SevenNet is a graph neural network interatomic potential package that supports efficient multi-GPU parallel molecular dynamics simulations. It provides pretrained universal potentials and enables large-scale simulations with excellent parallel scaling.

Scientific domain: Scalable ML potentials, parallel MD, graph neural networks
Target user community: Researchers needing scalable ML potentials for large systems

Theoretical Methods

  • Graph neural networks
  • E(3)-equivariant features
  • Multi-GPU parallelization
  • Universal potential training

Capabilities (CRITICAL)

  • Multi-GPU parallel MD
  • Pretrained universal models
  • LAMMPS integration
  • Large-scale simulations
  • Efficient scaling
  • ASE calculator

Key Strengths

Parallel Scaling:

  • Multi-GPU support
  • Efficient parallelization
  • Large systems
  • Good weak scaling

Pretrained Models:

  • Universal potentials
  • Ready to use
  • Fine-tuning support

Inputs & Outputs

  • Input formats:

    • ASE Atoms
    • LAMMPS data
  • Output data types:

    • Energies
    • Forces
    • Stresses
    • Model files

Interfaces & Ecosystem

  • LAMMPS: pair_style
  • ASE: Calculator
  • PyTorch: Backend

Advanced Features

  • Multi-GPU: Parallel inference
  • Universal models: Pretrained
  • LAMMPS: Large-scale MD
  • Fine-tuning: Transfer learning

Performance Characteristics

  • Excellent parallel scaling
  • Multi-GPU efficient
  • Fast inference
  • Good for large systems

Computational Cost

  • Pretrained: Ready to use
  • Inference: Fast, scales well
  • Training: GPU hours
  • Overall: Excellent scaling

Best Practices

  • Use multi-GPU for large systems
  • Start with pretrained models
  • Validate carefully
  • Fine-tune if needed

Limitations & Known Constraints

  • Requires multiple GPUs for best performance
  • Active development
  • Documentation evolving

Application Areas

  • Large-scale materials simulations
  • Parallel MD
  • Materials discovery
  • High-throughput screening

Verification & Sources

Primary sources:

  1. GitHub: https://github.com/MDIL-SNU/SevenNet
  2. Y. Park et al., J. Chem. Theory Comput. (2024)

Confidence: VERIFIED

Verification status: ✅ VERIFIED

  • Source code: OPEN (GitHub, GPL-3.0)

Related Tools in 6.3 Machine Learning Potentials