DPA-3.1-MPtrj

Version: v0.3 Added: 2025-06-05 Published: 2025-06-05 4.81M parameters Missing preds: 0
Repo Paper Docs DOI Files PR Checkpoint
Predictions
κSRMEGeometry OptimizationDiscovery

Convex hull distance prediction errors projected onto elements

1 H 0.22
2 He 0
3 Li 0.045
4 Be 0.0739
5 B 0.107
6 C 0.0844
7 N 0.128
8 O 0.135
9 F 0.14
10 Ne 0
11 Na 0.0467
12 Mg 0.0594
13 Al 0.0914
14 Si 0.101
15 P 0.0907
16 S 0.112
17 Cl 0.13
18 Ar 0
19 K 0.055
20 Ca 0.0577
21 Sc 0.0597
22 Ti 0.0653
23 V 0.087
24 Cr 0.113
25 Mn 0.14
26 Fe 0.124
27 Co 0.0798
28 Ni 0.0785
29 Cu 0.0659
30 Zn 0.0664
31 Ga 0.0847
32 Ge 0.0914
33 As 0.0897
34 Se 0.126
35 Br 0.113
36 Kr 0
37 Rb 0.0574
38 Sr 0.0543
39 Y 0.0653
40 Zr 0.0701
41 Nb 0.0885
42 Mo 0.0775
43 Tc 0.0694
44 Ru 0.101
45 Rh 0.0909
46 Pd 0.0814
47 Ag 0.0537
48 Cd 0.0572
49 In 0.0856
50 Sn 0.0873
51 Sb 0.0824
52 Te 0.143
53 I 0.101
54 Xe 0.02
55 Cs 0.0586
56 Ba 0.0586
57 La 0.0583
58 Ce 0.0579
59 Pr 0.0557
60 Nd 0.0542
61 Pm 0.0469
62 Sm 0.0563
63 Eu 0.0774
64 Gd 0.063
65 Tb 0.0588
66 Dy 0.0585
67 Ho 0.0568
68 Er 0.0647
69 Tm 0.0596
70 Yb 0.056
71 Lu 0.0572
72 Hf 0.0757
73 Ta 0.115
74 W 0.0814
75 Re 0.0777
76 Os 0.102
77 Ir 0.111
78 Pt 0.0983
79 Au 0.102
80 Hg 0.0559
81 Tl 0.0597
82 Pb 0.0831
83 Bi 0.0735
84 Po 0
85 At 0
86 Rn 0
87 Fr 0
88 Ra 0
89 Ac 0.0504
90 Th 0.0864
91 Pa 0.112
92 U 0.0927
93 Np 0.157
94 Pu 0.401
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. Duo Zhang  AI for Science Institute, Beijing  AI for Science Institute, Beijing logo  
  2. Anyang Peng  AI for Science Institute, Beijing  AI for Science Institute, Beijing logo  
  3. Chun Cai  AI for Science Institute, Beijing  AI for Science Institute, Beijing logo  
  4. Linfeng Zhang  AI for Science Institute, Beijing; DP Technology  
  5. Han Wang  Beijing Institute of Applied Physics and Computational Mathematics (IAPCM)  

Trained By

  1. Duo Zhang (AI for Science Institute, Beijing)

Model Info

  • Model Version v0.3
  • 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

DPA3 is an advanced interatomic potential leveraging the message passing architecture, implemented within the DeePMD-kit framework, available on GitHub. Designed as a large atomic model (LAM), DPA3 is tailored to integrate and simultaneously train on datasets from various disciplines, encompassing diverse chemical and materials systems across different research domains. Its model design ensures exceptional fitting accuracy and robust generalization both within and beyond the training domain. Furthermore, DPA3 maintains energy conservation and respects the physical symmetries of the potential energy surface, making it a dependable tool for a wide range of scientific applications.

Hyperparameters

  • max_force: 0.05
  • max_steps: 500
  • ase_optimizer: "FIRE"
  • cell_filter: "ExpCellFilter"
  • n_layers: 24
  • e_rcut: 6
  • a_rcut: 4.5
  • n_dim: 128
  • e_dim: 64
  • a_dim: 32
  • optimizer: "AdamW"
  • round1: {"loss":"MSE","loss_weights":{"energy":"0.2 -> 20","force":"100 -> 20","virial":"0.02 -> 1"},"initial_learning_rate":0.001,"learning_rate_schedule":"ExpLR - start_lr=0.001, decay_steps=5000, stop_lr=0.00001","training_steps":2000000}
  • round2: {"loss":"Huber","loss_weights":{"energy":15,"force":1,"virial":2.5},"initial_learning_rate":0.0002,"learning_rate_schedule":"ExpLR - start_lr=0.0002, decay_steps=5000, stop_lr=0.00001","training_steps":1000000}
  • batch_size: 64
  • epochs: 120
  • graph_construction_radius: 6
  • max_neighbors: null

Dependencies