BAM-MP-DeNS

Version: 1.0.0 Added: 2026-10-18 6.47M parameters

Model overview

Interatomic potential: Energy with gradient-based forces and stress.

CPS 0.672 combines discovery, geometry optimization, and phonons. Inspect individual task results below.

Code license
MIT
Checkpoint license
MIT

Licenses are as declared by the submitter. Training dataset terms are listed separately below.

Click a task score to show this model's diagnostics below.

Model info

Specifications

Architecture
gnn
Outputs
Energy with gradient-based forces and stress EFSG
Openness
Open source, open data OSOD

Discovery protocol

Training
structure to energy, force, stress S2EFS
Evaluation
initial structure to relaxed energy with structure relaxation IS2RE-SR

Training set

MPtrj: 1.58M structures from 146k materials · MIT

Model authors

  1. Hyun Gyu Park  Department of Energy Science, Sungkyunkwan University  
  2. Won Woong Choi  Department of Energy Science, Sungkyunkwan University  
  3. Seungwon Kim  Department of Energy Science, Sungkyunkwan University  
  4. Kwanghyeon Jo  Department of Energy Science, Sungkyunkwan University  
  5. Jooyeon Song  Department of Energy Science, Sungkyunkwan University  
  6. Wonho Lee  Department of Energy Science, Sungkyunkwan University  
  7. Gi Beom Sim  Department of Energy Science, Sungkyunkwan University  
  8. Tae Hyeon Park  Department of Energy Science, Sungkyunkwan University  
  9. Soohaeng Yoo Willow  Department of Energy Science, Sungkyunkwan University  
  10. Chang Woo Myung  Department of Energy Science, Sungkyunkwan University  

Trained by

  1. Hyun Gyu Park  Department of Energy Science, Sungkyunkwan University  

Description

BAM-MP-DeNS is a universal interatomic potential from the BAM-torch BAM_DeNS branch (RACE architecture), trained on MPtrj with denoising non-equilibrium structures (DeNS) as an auxiliary task. It is evaluated here on WBM IS2RE-SR discovery and geometry optimization, PhononDB kappa-103, DynaMat MD, and diatomic curves.

Hyperparams

Dependencies (6)

Run this model

Use the submitted environment with the upstream instructions.

From a repository checkout, run uv run models/run_kappa.py --list-models to check checkpoint requirements. If this model requires one, add --checkpoint /path/to/checkpoint below. Then execute the printed command for a phonon smoke test in the model's isolated environment.

uv run models/run_kappa.py --model bam-mp-dens --print-cmd --dry-run