model_name: EquiformerV3+DeNS-OAM
model_key: equiformer_v3_oam
model_version: v2026.04.07
date_added: '2026-04-07'
date_published: '2026-04-07'
authors:
  - name: Yi-Lun Liao
    affiliation: Massachusetts Institute of Technology
    email: ylliao@mit.edu
    github: https://github.com/yilunliao
    corresponding: true
  - name: Alexander J. Hoffman
    affiliation: Mirror Physics
  - name: Sabrina C. Shen
    affiliation: Mirror Physics
  - name: Sam Walton Norwood
    affiliation: Mirror Physics
    email: sam@mirrorphysics.com
    corresponding: true
  - name: Tess Smidt
    affiliation: Massachusetts Institute of Technology
    email: tsmidt@mit.edu
    url: https://blondegeek.github.io/
    orcid: https://orcid.org/0000-0001-5581-5344
    corresponding: true


trained_by:
  - name: Yi-Lun Liao
    affiliation: Massachusetts Institute of Technology
    email: ylliao@mit.edu


repo: https://github.com/atomicarchitects/equiformer_v3
doi: https://doi.org/10.48550/arXiv.2604.09130
paper: https://arxiv.org/abs/2604.09130
url: https://github.com/atomicarchitects/equiformer_v3
pr_url: https://github.com/janosh/matbench-discovery/pull/320
checkpoint_url: https://github.com/atomicarchitects/equiformer_v3


license:
  code: MIT
  code_url: https://github.com/atomicarchitects/equiformer_v3
  checkpoint: MIT
  checkpoint_url: https://github.com/atomicarchitects/equiformer_v3


openness: OSOD
trained_for_benchmark: true
train_task: S2EFS
test_task: IS2RE-SR
targets: EFS_G
model_type: UIP
model_params: 30_333_409
n_estimators: 1


training_set: [OMat24, MPtrj, sAlex]


training_cost:
  Nvidia H100 GPUs: {amount: 32, hours: 177.3125}


hyperparams:
  max_force: 0.02
  max_steps: 500
  ase_optimizer: FIRE
  cell_filter: FrechetCellFilter
  graph_construction_radius: 6.0
  max_neighbors: .inf

  num_blocks: 7
  lmax: 4
  mmax: 2
  embedding_dimension: 128
  attention_value_hidden_size: 32
  num_attention_heads: 8
  attention_alpha_hidden_size: 64
  attention_value_size_per_head: 16
  ffn_hidden_size_after_swiglu_s2_activation: 512
  grid_resolution_attention: (14, 8)
  grid_resolution_ffn: (14, 14)

  DeNS_probability: 0.5
  DeNS_coefficient: 1
  DeNS_noise_std: 0.025
  DeNS_corruption_ratio: 0.5
  DeNS_max_force_norm: 2.5


notes:
  Description: EquiformerV3 is the third generation of SE(3)-equivariant graph attention Transformers. Please refer to our paper for more details.


requirements:
  ase: 3.25.0
  ase-db-backends: 0.10.0
  e3nn: 0.5.6
  tqdm: 4.67.1
  wandb: 0.21.0
  pyyaml: 6.0.2
  torch: 2.7.1


metrics:
  phonons:
    kappa_103:
      pred_file: models/equiformer_v3/equiformer-v3-oam/2026-04-07-kappa-103-FIRE-dist=0.03-fmax=1e-4-symprec=1e-5.json.gz
      pred_file_url: https://figshare.com/files/63518859
      κ_SRME: 0.1178
      κ_SRE: 0.0495
  geo_opt:
    pred_file: models/equiformer_v3/equiformer-v3-oam/2026-04-07-wbm-IS2RE-FIRE.json.gz
    pred_file_url: https://figshare.com/files/63518865
    struct_col: structure
    symprec=1e-5:
      rmsd: 0.0595 # unitless
      n_sym_ops_mae: 4.0648 # unitless
      symmetry_decrease: 0.4554 # fraction
      symmetry_match: 0.3711 # fraction
      symmetry_increase: 0.1299 # fraction
      n_structures: 256963 # count
      analysis_file: models/equiformer_v3/equiformer-v3-oam/2026-04-07-wbm-IS2RE-FIRE-symprec=1e-5-moyo=0.7.9.csv.gz
      analysis_file_url: https://figshare.com/files/63520848
    symprec=1e-2:
      rmsd: 0.0595 # unitless
      n_sym_ops_mae: 1.9824 # unitless
      symmetry_decrease: 0.129 # fraction
      symmetry_match: 0.7424 # fraction
      symmetry_increase: 0.1137 # fraction
      n_structures: 256963 # count
      analysis_file: models/equiformer_v3/equiformer-v3-oam/2026-04-07-wbm-IS2RE-FIRE-symprec=1e-2-moyo=0.7.9.csv.gz
      analysis_file_url: https://figshare.com/files/63520845
  discovery:
    pred_file: models/equiformer_v3/equiformer-v3-oam/2026-04-07-wbm-IS2RE.csv.gz
    pred_file_url: https://figshare.com/files/63518862
    pred_col: e_form_per_atom_fairchem
    full_test_set:
      F1: 0.912 # fraction
      DAF: 5.298 # dimensionless
      Precision: 0.909 # fraction
      Recall: 0.916 # fraction
      Accuracy: 0.97 # fraction
      TPR: 0.916 # fraction
      FPR: 0.019 # fraction
      TNR: 0.981 # fraction
      FNR: 0.084 # fraction
      TP: 40369.0 # count
      FP: 4035.0 # count
      TN: 208836.0 # count
      FN: 3723.0 # count
      MAE: 0.017 # eV/atom
      RMSE: 0.067 # eV/atom
      R2: 0.862 # dimensionless
      missing_preds: 2 # count
    unique_prototypes:
      F1: 0.931 # fraction
      DAF: 6.074 # dimensionless
      Precision: 0.928 # fraction
      Recall: 0.933 # fraction
      Accuracy: 0.978 # fraction
      TPR: 0.933 # fraction
      FPR: 0.013 # fraction
      TNR: 0.987 # fraction
      FNR: 0.067 # fraction
      TP: 31139.0 # count
      FP: 2398.0 # count
      TN: 179716.0 # count
      FN: 2235.0 # count
      MAE: 0.018 # eV/atom
      RMSE: 0.067 # eV/atom
      R2: 0.868 # dimensionless
      missing_preds: 0 # count
    most_stable_10k:
      F1: 0.986 # fraction
      DAF: 6.357 # dimensionless
      Precision: 0.972 # fraction
      Recall: 1.0 # fraction
      Accuracy: 0.972 # fraction
      TPR: 1.0 # fraction
      FPR: 1.0 # fraction
      TNR: 0.0 # fraction
      FNR: 0.0 # fraction
      TP: 9718.0 # count
      FP: 282.0 # count
      TN: 0.0 # count
      FN: 0.0 # count
      MAE: 0.015 # eV/atom
      RMSE: 0.067 # eV/atom
      R2: 0.902 # dimensionless
      missing_preds: 0 # count