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 0.00
3 Li 0.04
4 Be 0.07
5 B 0.11
6 C 0.08
7 N 0.13
8 O 0.13
9 F 0.14
10 Ne 0.00
11 Na 0.05
12 Mg 0.06
13 Al 0.09
14 Si 0.10
15 P 0.09
16 S 0.11
17 Cl 0.13
18 Ar 0.00
19 K 0.06
20 Ca 0.06
21 Sc 0.06
22 Ti 0.07
23 V 0.09
24 Cr 0.11
25 Mn 0.14
26 Fe 0.12
27 Co 0.08
28 Ni 0.08
29 Cu 0.07
30 Zn 0.07
31 Ga 0.08
32 Ge 0.09
33 As 0.09
34 Se 0.13
35 Br 0.11
36 Kr 0.00
37 Rb 0.06
38 Sr 0.05
39 Y 0.07
40 Zr 0.07
41 Nb 0.09
42 Mo 0.08
43 Tc 0.07
44 Ru 0.10
45 Rh 0.09
46 Pd 0.08
47 Ag 0.05
48 Cd 0.06
49 In 0.09
50 Sn 0.09
51 Sb 0.08
52 Te 0.14
53 I 0.10
54 Xe 0.02
55 Cs 0.06
56 Ba 0.06
57 La 0.06
58 Ce 0.06
59 Pr 0.06
60 Nd 0.05
61 Pm 0.05
62 Sm 0.06
63 Eu 0.08
64 Gd 0.06
65 Tb 0.06
66 Dy 0.06
67 Ho 0.06
68 Er 0.06
69 Tm 0.06
70 Yb 0.06
71 Lu 0.06
72 Hf 0.08
73 Ta 0.12
74 W 0.08
75 Re 0.08
76 Os 0.10
77 Ir 0.11
78 Pt 0.10
79 Au 0.10
80 Hg 0.06
81 Tl 0.06
82 Pb 0.08
83 Bi 0.07
84 Po 0.00
85 At 0.00
86 Rn 0.00
87 Fr 0.00
88 Ra 0.00
89 Ac 0.05
90 Th 0.09
91 Pa 0.11
92 U 0.09
93 Np 0.16
94 Pu 0.40
95 Am 0.00
96 Cm 0.00
97 Bk 0.00
98 Cf 0.00
99 Es 0.00
100 Fm 0.00
101 Md 0.00
102 No 0.00
103 Lr 0.00
104 Rf 0.00
105 Db 0.00
106 Sg 0.00
107 Bh 0.00
108 Hs 0.00
109 Mt 0.00
110 Ds 0.00
111 Rg 0.00
112 Cn 0.00
113 Nh 0.00
114 Fl 0.00
115 Mc 0.00
116 Lv 0.00
117 Ts 0.00
118 Og 0.00
57-71 La-Lu Lanthanides
89-103 Ac-Lr Actinides

Model Authors

  1. Duo Zhang  AI for Science Institute, Beijing    
  2. Anyang Peng  AI for Science Institute, Beijing    
  3. Chun Cai  AI for Science Institute, Beijing    
  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