Allegro-MP-L

Version: 0.1 Added: 2025-09-08 Published: 2025-08-28 18.7M parameters Missing preds: 0
Repo Paper Docs DOI Files PR Checkpoint
Predictions
κSRMEGeometry OptimizationDiscovery

Convex hull distance prediction errors projected onto elements

1 H 0.29
2 He 0
3 Li 0.0563
4 Be 0.093
5 B 0.121
6 C 0.115
7 N 0.173
8 O 0.235
9 F 0.306
10 Ne 0
11 Na 0.0739
12 Mg 0.0648
13 Al 0.099
14 Si 0.101
15 P 0.112
16 S 0.187
17 Cl 0.263
18 Ar 0
19 K 0.0888
20 Ca 0.0737
21 Sc 0.0772
22 Ti 0.0828
23 V 0.0969
24 Cr 0.122
25 Mn 0.146
26 Fe 0.13
27 Co 0.0823
28 Ni 0.0808
29 Cu 0.0694
30 Zn 0.0737
31 Ga 0.0908
32 Ge 0.0946
33 As 0.103
34 Se 0.184
35 Br 0.226
36 Kr 0
37 Rb 0.0917
38 Sr 0.0722
39 Y 0.0856
40 Zr 0.0924
41 Nb 0.103
42 Mo 0.0948
43 Tc 0.0795
44 Ru 0.113
45 Rh 0.0942
46 Pd 0.088
47 Ag 0.0632
48 Cd 0.0617
49 In 0.0964
50 Sn 0.0957
51 Sb 0.0918
52 Te 0.177
53 I 0.179
54 Xe 0.004
55 Cs 0.0976
56 Ba 0.0772
57 La 0.0722
58 Ce 0.066
59 Pr 0.0704
60 Nd 0.0705
61 Pm 0.0527
62 Sm 0.0717
63 Eu 0.0825
64 Gd 0.0824
65 Tb 0.0756
66 Dy 0.0798
67 Ho 0.0763
68 Er 0.0796
69 Tm 0.0809
70 Yb 0.0737
71 Lu 0.0762
72 Hf 0.098
73 Ta 0.14
74 W 0.0931
75 Re 0.0896
76 Os 0.11
77 Ir 0.117
78 Pt 0.0972
79 Au 0.101
80 Hg 0.0649
81 Tl 0.0676
82 Pb 0.096
83 Bi 0.0843
84 Po 0
85 At 0
86 Rn 0
87 Fr 0
88 Ra 0
89 Ac 0.0584
90 Th 0.1
91 Pa 0.112
92 U 0.103
93 Np 0.15
94 Pu 0.333
95 Am 0
96 Cm 0
97 Bk 0
98 Cf 0
99 Es 0
100 Fm 0
101 Md 0
102 No 0
103 Lr 0
104 Rf 0
105 Db 0
106 Sg 0
107 Bh 0
108 Hs 0
109 Mt 0
110 Ds 0
111 Rg 0
112 Cn 0
113 Nh 0
114 Fl 0
115 Mc 0
116 Lv 0
117 Ts 0
118 Og 0
57-71 La-Lu Lanthanides
89-103 Ac-Lr Actinides

Model Authors

  1. Seán R. Kavanagh  Center for the Environment, Harvard University & MIR Group, Harvard University    
  2. Chuin Wei Tan  MIR Group, Harvard University  MIR Group, Harvard University logo
  3. Albert Musaelian  MIR Group, Harvard University & Mirian Technologies
  4. William C. Witt  MIR Group, Harvard University  MIR Group, Harvard University logo
  5. Gabriel de Miranda Nascimento  MIR Group, Harvard University & MIT
  6. Ulrik Unneberg  MIR Group, Harvard University & MIT
  7. Marc L. Descoteaux  MIR Group, Harvard University  MIR Group, Harvard University logo
  8. Boris Kozinsky  MIR Group, Harvard University  MIR Group, Harvard University logo

Trained By

  1. Seán R. Kavanagh (Center for the Environment, Harvard University & MIR Group, Harvard University)

Model Info

  • Model Version 0.1
  • Model Type UIP
  • Targets EFSG
  • Openness OSOD
  • Train Task S2EFS
  • Test Task IS2RE-SR
  • Trained for Benchmark Yes

Training Set

MPtrj: 1.58M structures from 146k materials

Description

Large 'compliant' Allegro foundation potential; see https://www.nequip.net/models/mir-group/Allegro-MP-L:0.1 for details and https://arxiv.org/abs/2504.16068 for model/training infrastructure.

Steps

Single training run on MPtrj with specified parameters.

Hyperparameters

  • max_force: 0.05
  • max_steps: 500
  • ase_optimizer: "GOQN"
  • cell_filter: "FrechetCellFilter"
  • optimizer: "AdamW"
  • weight_decay: 0.001
  • graph_construction_radius: 6
  • sph_harmonics_l_max: 3
  • n_layers: 5
  • n_features: 96
  • parity: false
  • zbl_potential: false
  • allegro_mlp_depth: 3
  • allegro_mlp_width: 1024
  • tensor_path_channel_coupling: true
  • polynomial_cutoff: 8
  • n_radial_bessel_basis: 12
  • loss: "Huber - delta=0.01 for energy, delta=0.1 for stress, stratified delta (0.01, 0.007, 0.004, 0.001) for force"
  • loss_weights: {"energy":1,"force":5,"stress":0.01}
  • batch_size: 520
  • initial_learning_rate: 0.01
  • gradient_clip_val: 0.015
  • learning_rate_schedule: "ReduceLROnPlateau - factor=0.5, patience=250, min_lr=1e-6"
  • epochs: 250
  • max_neighbors: null

Dependencies