model_name: Nequix MP PFT model_key: nequix-mp-1-pft model_version: '0.1.0' date_added: '2026-01-08' date_published: '2026-01-08' authors: - name: Teddy Koker affiliation: Massachusetts Institute of Technology email: tekoker@mit.edu url: https://teddykoker.com orcid: https://orcid.org/0000-0001-8861-9788 - name: Abhijeet Gangan affiliation: University of California, Los Angeles email: abhijeetgangan@g.ucla.edu orcid: https://orcid.org/0000-0002-8937-7984 - name: Mit Kotak affiliation: Massachusetts Institute of Technology email: mkotak@mit.edu orcid: https://orcid.org/0000-0002-6999-1055 - name: Jaime Marian affiliation: University of California, Los Angeles email: jmarian@g.ucla.edu orcid: https://orcid.org/0000-0001-9000-3405 - 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 trained_by: - name: Teddy Koker affiliation: Massachusetts Institute of Technology email: tekoker@mit.edu url: https://teddykoker.com orcid: https://orcid.org/0000-0001-8861-9788 repo: https://github.com/atomicarchitects/nequix url: https://github.com/atomicarchitects/nequix doi: https://doi.org/10.48550/arXiv.2601.07742 paper: https://arxiv.org/abs/2601.07742 pr_url: https://github.com/janosh/matbench-discovery/pull/304 checkpoint_url: https://figshare.com/files/60965527 openness: OSOD train_task: S2EFS test_task: IS2RE-SR targets: EFSH_G model_type: UIP model_params: 707_658 trained_for_benchmark: true n_estimators: 1 license: code: MIT code_url: https://github.com/atomicarchitects/nequix/blob/main/LICENSE checkpoint: CC-BY-4.0 checkpoint_url: https://creativecommons.org/licenses/by/4.0/ hyperparams: max_force: 0.02 max_steps: 500 ase_optimizer: FIRE cell_filter: FrechetCellFilter hidden_irreps: 128x0e+64x1o+32x2e+32x3o graph_construction_radius: 6.0 max_neighbors: .inf lmax: 3 radial_basis_size: 8 radial_mlp_size: 64 radial_mlp_layers: 2 radial_polynomial_p: 6.0 layer_norm: true optimizer: adamw learning_rate: 0.0001 weight_decay: 0.001 grad_clip_norm: 100 batch_size: 16 n_epochs: 100 loss: energy: mae force: l2 stress: mae hessian: mae loss_weights: energy: 0.0 force: 20.0 stress: 5.0 hessian: 100.0 cotrain_loss_weights: energy: 500.0 force: 200.0 stress: 50.0 ema_decay: 0.999 training_cost: Nvidia A100 GPUs: { amount: 1, hours: 140 } training_set: [MPtrj, MDR PBE Phonons in MPtrj] notes: Description: Nequix MP PFT is a fine-tuned version of Nequix MP 1, which was fine-tuned using phonon fine-tuning (PFT) on a subset of the MDR Phonon PBE dataset that is contained within MPtrj. Reproduction: | git clone https://github.com/atomicarchitects/nequix.git && cd nequix uv sync --extra pft bash data/download_pbe_mdr_preprocessed.sh uv run nequix/pft/train.py configs/nequix-mp-1-pft.yml requirements: jax: 0.6.2 e3nn-jax: 0.20.7 jraph: 0.0.6.dev0 equinox: 0.11.11 optax: 0.2.5 ase: 3.24.0 tqdm: 4.67.1 wandb: 0.19.11 pyyaml: 6.0.2 matscipy: 1.1.1 h5py: 3.14.0 wandb-osh: 1.2.2 cloudpickle: 3.1.1 metrics: phonons: kappa_103: pred_file: models/nequix/nequix-mp-1-pft/2026-01-12-kappa-103-FIRE-dist=0.03-fmax=1e-4-symprec=1e-5.json.gz pred_file_url: https://figshare.com/files/61019518 κ_SRME: 0.3066 κ_SRE: 0.1689 geo_opt: pred_file: models/nequix/nequix-mp-1-pft/2026-01-12-wbm-IS2RE-FIRE.jsonl.gz pred_file_url: https://figshare.com/files/61019686 struct_col: nequix-mp-pft-1yqa7d5t-200epochs_structure symprec=1e-2: rmsd: 0.0866 # unitless n_sym_ops_mae: 1.7984 # unitless symmetry_decrease: 0.054 # fraction symmetry_match: 0.8109 # fraction symmetry_increase: 0.1276 # fraction n_structures: 256963 # count analysis_file: models/nequix/nequix-mp-1-pft/2026-01-12-wbm-IS2RE-FIRE-symprec=1e-2-moyo=0.7.4.csv.gz analysis_file_url: https://figshare.com/files/61021306 symprec=1e-5: rmsd: 0.0866 # unitless n_sym_ops_mae: 2.0578 # unitless symmetry_decrease: 0.0602 # fraction symmetry_match: 0.6976 # fraction symmetry_increase: 0.2362 # fraction n_structures: 256963 # count analysis_file: models/nequix/nequix-mp-1-pft/2026-01-12-wbm-IS2RE-FIRE-symprec=1e-5-moyo=0.7.4.csv.gz analysis_file_url: https://figshare.com/files/61021309 discovery: pred_file: models/nequix/nequix-mp-1-pft/2026-01-12-wbm-IS2RE.csv.gz pred_file_url: https://figshare.com/files/61019521 pred_col: e_form_per_atom_nequix full_test_set: F1: 0.74 # fraction DAF: 3.963 # dimensionless Precision: 0.68 # fraction Recall: 0.811 # fraction Accuracy: 0.902 # fraction TPR: 0.811 # fraction FPR: 0.079 # fraction TNR: 0.921 # fraction FNR: 0.189 # fraction TP: 35757.0 # count FP: 16821.0 # count TN: 196050.0 # count FN: 8335.0 # count MAE: 0.042 # eV/atom RMSE: 0.084 # eV/atom R2: 0.783 # dimensionless missing_preds: 2 # count unique_prototypes: F1: 0.748 # fraction DAF: 4.479 # dimensionless Precision: 0.685 # fraction Recall: 0.825 # fraction Accuracy: 0.914 # fraction TPR: 0.825 # fraction FPR: 0.07 # fraction TNR: 0.93 # fraction FNR: 0.175 # fraction TP: 27533.0 # count FP: 12676.0 # count TN: 169438.0 # count FN: 5841.0 # count MAE: 0.044 # eV/atom RMSE: 0.085 # eV/atom R2: 0.784 # dimensionless missing_preds: 0 # count most_stable_10k: F1: 0.953 # fraction DAF: 5.95 # dimensionless Precision: 0.91 # fraction Recall: 1.0 # fraction Accuracy: 0.91 # fraction TPR: 1.0 # fraction FPR: 1.0 # fraction TNR: 0.0 # fraction FNR: 0.0 # fraction TP: 9096.0 # count FP: 904.0 # count TN: 0.0 # count FN: 0.0 # count MAE: 0.051 # eV/atom RMSE: 0.114 # eV/atom R2: 0.744 # dimensionless missing_preds: 0 # count