MatterSim v1 5M

Version: v1.0.0 Added: 2024-12-16 Published: 2024-05-08 4.55M parameters
Leaderboard ranks CPS #24 /41Discovery F1 #22 /52Geo Opt RMSD #23 /42Phonons κSRME #31 /41MD CMDS #11 /34Diatomics CDS #8 /33

Discovery: energy and convex hull diagnostics

Missing preds: 0
Loading formation energy parity data...

Per-element convex hull distance errors

1 H 0.15
2 He Helium
3 Li 0.03
4 Be 0.04
5 B 0.06
6 C 0.05
7 N 0.07
8 O 0.11
9 F 0.11
10 Ne Neon
11 Na 0.03
12 Mg 0.05
13 Al 0.05
14 Si 0.06
15 P 0.06
16 S 0.08
17 Cl 0.10
18 Ar Argon
19 K 0.04
20 Ca 0.04
21 Sc 0.04
22 Ti 0.04
23 V 0.06
24 Cr 0.08
25 Mn 0.11
26 Fe 0.09
27 Co 0.05
28 Ni 0.05
29 Cu 0.05
30 Zn 0.04
31 Ga 0.05
32 Ge 0.07
33 As 0.06
34 Se 0.10
35 Br 0.09
36 Kr Krypton
37 Rb 0.04
38 Sr 0.04
39 Y 0.04
40 Zr 0.04
41 Nb 0.05
42 Mo 0.05
43 Tc 0.03
44 Ru 0.05
45 Rh 0.05
46 Pd 0.05
47 Ag 0.04
48 Cd 0.04
49 In 0.06
50 Sn 0.05
51 Sb 0.06
52 Te 0.13
53 I 0.08
54 Xe 0.02
55 Cs 0.04
56 Ba 0.04
57 La 0.04
58 Ce 0.04
59 Pr 0.04
60 Nd 0.03
61 Pm 0.03
62 Sm 0.04
63 Eu 0.07
64 Gd 0.06
65 Tb 0.03
66 Dy 0.03
67 Ho 0.03
68 Er 0.03
69 Tm 0.03
70 Yb 0.04
71 Lu 0.03
72 Hf 0.05
73 Ta 0.07
74 W 0.05
75 Re 0.04
76 Os 0.05
77 Ir 0.06
78 Pt 0.06
79 Au 0.07
80 Hg 0.03
81 Tl 0.04
82 Pb 0.06
83 Bi 0.05
84 Po Polonium
85 At Astatine
86 Rn Radon
87 Fr Francium
88 Ra Radium
89 Ac 0.03
90 Th 0.05
91 Pa 0.06
92 U 0.06
93 Np 0.09
94 Pu 0.20
95 Am Americium
96 Cm Curium
97 Bk Berkelium
98 Cf Californium
99 Es Einsteinium
100 Fm Fermium
101 Md Mendelevium
102 No Nobelium
103 Lr Lawrencium
104 Rf Rutherfordium
105 Db Dubnium
106 Sg Seaborgium
107 Bh Bohrium
108 Hs Hassium
109 Mt Meitnerium
110 Ds Darmstadtium
111 Rg Roentgenium
112 Cn Copernicum
113 Nh Nihonium
114 Fl Flerovium
115 Mc Moscovium
116 Lv Livermorium
117 Ts Tennessine
118 Og Oganesson
57-71 La-Lu Lanthanides
89-103 Ac-Lr Actinides

ML vs DFT Lattice Thermal Conductivity

Loading κ parity data...

Model Authors

  1. Han Yang Microsoft Research AI for Science  
  2. Chenxi Hu Microsoft Research AI for Science  
  3. Yichi Zhou Microsoft Research AI for Science  
  4. Xixian Liu Microsoft Research AI for Science  
  5. Yu Shi Microsoft Research AI for Science  
  6. Jielan Li Microsoft Research AI for Science  
  7. Guanzhi Li Microsoft Research AI for Science  
  8. Zekun Chen Microsoft Research AI for Science  
  9. Shuizhou Chen Microsoft Research AI for Science  
  10. Claudio Zeni Microsoft Research AI for Science  
  11. Matthew Horton Microsoft Research AI for Science  
  12. Robert Pinsler Microsoft Research AI for Science  
  13. Andrew Fowler Microsoft Research AI for Science  
  14. Daniel Zügner Microsoft Research AI for Science  
  15. Tian Xie Microsoft Research AI for Science  
  16. Jake Smith Microsoft Research AI for Science  
  17. Lixin Sun Microsoft Research AI for Science  
  18. Qian Wang Microsoft Research AI for Science  
  19. Lingyu Kong Microsoft Research AI for Science  
  20. Chang Liu Microsoft Research AI for Science  
  21. Hongxia Hao Microsoft Research AI for Science  
  22. Ziheng Lu Microsoft Research AI for Science  

Model Info

  • Version v1.0.0
  • Role Interatomic potential
  • Architecture gnn
  • Targets EFSG
  • Openness OSCD
  • Discovery Train Task S2EFS
  • Discovery Test Task IS2RE-SR

Training Set

MatterSim: 17M structures

description

This is an open source version of MatterSim V1 based on M3GNet architecture.

training

MatterSim was trained on a large, closed dataset covering diverse combinations of 89 elements across many temperatures and pressures.

tested_applications

  • Energy, force, stress prediction
  • Molecular dynamics simulations
  • Phonons
  • Mechanical properties
  • Free energy and phase diagrams
  • Materials discovery

training_data_sources

  • Materials Project
  • Alexandria dataset
  • newly generated structures and MD trajectories

Hyperparams

  • evaluation: {"ase_optimizer":"FIRE","cell_filter":"ExpCellFilter","max_steps":500,"max_force":0.02,"kappa":{"protocol":"phonondb-v1"}}
  • architecture: {"graph_construction_radius":5}
  • training: {"learning_rate":0.0005,"batch_size":128,"optimizer":"AdamW"}
  • upstream_config: {"seed":42,"units":256}

Dependencies