IANN: A Unified Framework for Equivariant InterAtomic Neural Networks with User-Friendly Implementation
Authors: Changzhi Ai, Xueping Qin, Dimosthenis Sokaras, Frank Abild-Pedersen, Johannes Voss
Organizations: Department of Chemical Engineering, Stanford University, Stanford, California 94305, United States · SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States · Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States · Center for Nanoscale Materials, Argonne National Laboratory, Lemont, Illinois, United States
Machine-learning interatomic potentials (MLIPs) now approach density functional theory (DFT) accuracy at a fraction of its cost, but the practical benefit is limited by software fragmentation: each state-of-the-art architecture is released as a separate codebase, with its own data format, configuration system, training loop and molecular-dynamics interface. Comparing two architectures on the same dataset, or moving a trained model into a production simulation, therefore requires substantial re-engineering. We introduce IANN (InterAtomic Neural Network), an open-source PyTorch framework that reimplements equivariant graph neural networks - PaiNN, NequIP, MACE, Allegro, EquiformerV2, EquiformerV3 and UMA - behind a unified data object, a unified trainer and a single deployment route, so that the architecture is a one-word change in a configuration dictionary. Forces, stresses and virials are obtained for every architecture from one analytic differentiation of the energy with respect to the edge vectors of the atomic graph, exact under periodic boundary conditions. Distributed training is configured automatically from the SLURM, PBS or MPI environment, so the same script runs unchanged from a workstation to a multi-node GPU allocation. Self-describing checkpoints serve directly as ASE calculators, as ensembles reporting uncertainties, or as TorchScript models run inside LAMMPS through bundled C++ pair styles with single- and multi-GPU inference. Trained on a common dataset of millions of structures, the architectures can be compared under identical conditions. We further release PaiNN foundation models trained on 41.9 million curated DFT structures from MPtrj, Alexandria, OMat24, MatPES, OC20, OC22, OC25, MPr2SCAN and MatPESr2SCAN, spanning the PBE, RPBE and r2SCAN functionals. Additionally, IANN provides an agent interface to facilitate integration with autonomous agents.
Figures & tables
Figure 1: Software structure of the IANN framework: five main layers with two auxiliary layers.
Figure 2: The underlying mechanism: a shared graph convention, a unified data object, and a shared derivative route. (a) A structure becomes a graph: atom i bonds to atom j within the cutoff rc . (b) The unified data object: AtomsData , which describes the graph and its associated data. (c) The architecture-independent forward pass that yields forces, stress and virial.
Functional
Source dataset
System type
Structures
Atoms (M)
Elem.
Ref.
PBE
MPtrj
Bulk, relaxations
1,579,091
49.2
89
[ 14 , 15 ]
sAlex
Bulk, relaxations
10,423,284
107.5
89
[ 44 , 6 ]
OMat24
Bulk, off-equilibrium
11,274,650
157.1
89
[ 6 ]
MatPES
Bulk, MD snapshots
434,092
3.9
89
[ 23 ]
OC22
Surfaces, oxides
8,199,197
654.3
57
[ 54 ]
PBE total
31,910,314
972.0
89
Table 1: Curated DFT databases grouped by exchange-correlation functional.
Figure 3: Elemental composition of three levels of curated DFT databases: number of structures containing each element on a logarithmic colour scale. Each element tile is divided into three stepped sub-cells from top left to bottom right, corresponding to PBE, RPBE, and r 2 SCAN.
Model
Parameters
EMAE (eV/atom)
FMAE (eV/Å)
PaiNN
600,577
0.041
0.050
NequIP
1,379,513
0.050
0.069
Allegro
567,624
0.040
0.066
MACE
1,429,376
0.054
0.069
EquiformerV2
773,089
0.040
0.059
EquiformerV3
943,249
0.040
0.046
Table 2: Validation accuracy and parameter counts for seven representative architectures trained via IANN on the curated MPtrj dataset. [ 5 ] Parameter count serves as the common metric of model capacity across disparate architectures; network depth, channel dimensions, and angular resolutions for each model are detailed in Table .
Functional
Model Name
Database
Structure Count
EMAE (eV/atom)
FMAE (eV/Å)
PBE
PBE-MPtrj
MPtrj
1,579,058
0.0517
0.0498
PBE-sAlex
sAlex
10,422,893
0.0408
0.0513
PBE-OMat24
OMat24
11,261,400
0.0261
0.1382
PBE-MatPES
MatPES
434,083
0.0463
0.0930
PBE-OC22
OC22
8,198,695
0.0381
0.0528
RPBE
RPBE-OC20
OC20
1,999,216
0.0263
0.0630
Table 3: PaiNN foundation models within IANN grouped by three levels of DFT exchange-correlation functional. [ 5 ] Each model is trained on a single source database. Errors here are evaluated across entire source databases rather than on validation sets.
Figure 4: The PaiNN foundation models trained on each DFT database. (a) Energy MAE against force MAE. (b) Number of structures that each was trained on. (c, e, g) Per-element energy MAE against atomic number. (d, f, h) Per-element force MAE against atomic number for the PBE, RPBE and r 2 SCAN functionals, respectively.
Figure 5: Benchmark performance of the RPBE-OC20 PaiNN foundation model on the AdsTraj. (a) Heatmap of energy MAE across transition metal surfaces (Sc to Au) and key catalytic adsorbates (C*, H*, O*, CO*, OH*, and HOCO*). (b) Energy MAE resolved across metals. Parity plots of adsorption energy under (c) zero-shot, (d) from scratch, and (e) fine-tuning.
Strategy
Training structures
Eads MAE (eV)
R2
Zero-shot (foundation model)
0
0.715
0.846
From scratch (train from scratch)
24,154
0.483
0.926
Fine-tuned (train on foundation model)
24,154
0.217
0.986
Table 4: Performance comparison across zero-shot, training from scratch, and fine-tuning.
Figure 6: Inference cost of the architectures via IANN. (a) Wall time per energy-and-force call against system size. (b) Peak GPU memory versus the number of atoms. (c) Force MAE against sustained throughput for each model. (d) NNP Speedups over DFT (VASP, RPBE). The shaded region is where the DFT reference is extrapolated.
Figure 7: Performance of the IANN LAMMPS interface on A100-80GB GPUs. (a) Time per MD step against the number of atoms for 1–8 GPUs. (b) Parallel efficiency against GPU count. (c) Throughput against the number of atoms. (d) Maximum number of atoms versus GPU count.
Appendix figures & tables60 assets
Supplementary material from the paper’s appendix.
Appendix
Key
Default
Meaning
Model
num_channels
128
Feature channels per node
num_layers
3
Message-passing layers
cutoff
5.5
Cutoff radius (Å)
Objective
forces_weight
0.9
Weight wF of the force term
Appendix
Table S1: Configuration keys of iann.trainer.Trainer and their default values.
Model
Channels
ℓmax
mmax
Layers
Parameters
Batch
LR
Clip
PaiNN ∗
128
1
1
3
600,577
16
1×10−4
10.0
NequIP
32
1
1
3
1,379,513
16
1×10−4
10.0
Allegro
128
1
1
4
567,624
16
1×10−4
10.0
MACE
48
1
1
3
1,429,376
16
1×10−4
10.0
EquiformerV2
16
4
2
3
773,089
8
1×10−4
10.0
EquiformerV3
32
3
2
3
943,249
16
1×10−4
10.0
Appendix
Table S2: Settings of the training of each architecture. ℓmax is the highest spherical harmonic degree the features carry and mmax the highest order kept at each degree; NequIP, MACE and Allegro build full O(3) irreps, so mmax=ℓmax , whereas the Equiformers truncate below ℓmax to cut the cost of the SO(2) convolution. ∗ Equivalent values: PaiNN exposes neither setting, but its Cartesian scalars and vectors are equivariant to ℓ≤1 with all orders present.
Dataset
Nstruct
Natom
Energy (eV atom -1 )
Force (eV Å -1 )
MAE
R2
MAE
R2
PBE
Alexandria
10,422,893
107,451,444
0.0408
0.9988
0.0513
0.9512
MatPES
434,083
3,865,201
0.0463
0.9984
0.0930
0.9583
MPtrj
1,579,058
49,229,839
0.0517
0.9980
0.0498
0.9767
OC22
8,198,695
654,300,266
0.0381
0.9986
0.0528
0.8373
Appendix
Table S3: Accuracy of the released PaiNN foundation models. Energies are per atom; forces are per Angstrom. Nstruct and Natom count the structures and atoms entering each fit.
Figure S1: Energy and force parity plots between the prediction of the PBE-MPtrj foundation model and the DFT reference.
Figure S2: Energy and force parity plots between the prediction of the PBE-sAlex foundation model and the DFT reference.
Figure S3: Energy and force parity plots between the prediction of the PBE-OMat24 foundation model and the DFT reference.
Figure S4: Energy and force parity plots between the prediction of the PBE-MatPES foundation model and the DFT reference.
Figure S5: Energy and force parity plots between the prediction of the PBE-OC22 foundation model and the DFT reference.
Figure S6: Energy and force parity plots between the prediction of the RPBE-OC20 foundation model and the DFT reference.
Figure S7: Energy and force parity plots between the prediction of the RPBE-OC25 foundation model and the DFT reference.
Figure S8: Energy and force parity plots between the prediction of the r 2 SCAN-MPtrj foundation model and the DFT reference.
Figure S9: Energy and force parity plots between the prediction of the r 2 SCAN-MatPES foundation model and the DFT reference.
Test error
Strategy
Structures
EMAE
FMAE
Eads,MAE
(meV atom -1 )
(meV Å -1 )
(eV)
Zero-shot
0
25.0
67.2
0.715
From scratch
86
2283.6
565.7
71.1
1,023
794.0
181.4
28.2
2,119
524.6
223.7
18.1
Appendix
Table S4: Data efficiency of fine-tuning on the AdsTraj dataset.
Figure S10: Cost of training measured on the same V100 as inference. (a) Wall time per optimizer step against batch size, in 108-atom cells. (b) Training throughput in structures per second, and where it saturates. (c) Peak GPU memory versus batch size. (d) One training step against one inference call on the same cell. The corresponding numbers are in Table S5 .
Architecture
Ba
step b
struct./s
GPU-h per
mem
step at B=16c
vs
(ms)
106 steps
(GiB)
(ms)
inference d
PaiNN
16
86.0
186
23.9
3.8
86.0
2.8 ×
NequIP
16
110.1
145
30.6
2.0
110.1
2.2 ×
Allegro
16
935.4
17
259.8
10.9
935.4
2.8 ×
MACE
16
91.3
175
25.4
2.7
91.3
2.5 ×
EquiformerV2
8
824.9
10
229.1
19.9
1649.8
2.4 ×
Appendix
Table S5: Cost and corresponding details of all architectures trained on Tesla V100 GPU.
Architecture
Params
Floor a
Crossover b
N∗c
Max Nd
Mem @ 103
MD e
EMAE f
FMAE f
(ms)
( μ s/atom)
(GiB)
(ns/day)
(meV/at.)
(meV/Å)
PaiNN
600,581
9.0
15.9
564
10,976
1.36
3.00
41
50
NequIP
1,383,079
24.3
18.5
1,316
10,976
0.92
2.03
50
69
Allegro
570,062
19.1
180.7
106
4,000
5.72
0.76
40
66
MACE
1,435,686
20.8
15.0
1,391
10,976
0.76
2.23
54
69
EquiformerV2
773,093
114.6
306.7
374
864
19.77
0.36
40
59
Appendix
Table S6: Inference cost and accuracy of different architectures on one Tesla V100-PCIE-32GB, float32, 5.5 Å cutoff for energy+forces on bulk fcc Pt supercells.
Architecture
Floor E
Crossover E
Floor E+F
Crossover E+F
Force cost, a
Force cost, a
Mem Eb
Mem E+Fb
(ms)
( μ s/atom)
(ms)
( μ s/atom)
Floor
Crossover
(GiB)
(GiB)
PaiNN
3.1
6.9
9.0
15.9
2.90 ×
2.30 ×
0.33
1.36
NequIP
6.7
5.5
24.3
18.5
3.61 ×
3.33 ×
0.29
0.92
Allegro
4.5
54.3
19.1
180.7
4.28 ×
3.33 ×
1.07
5.72
MACE
8.2
5.5
20.8
15.0
2.55 ×
2.72 ×
0.21
0.76
EquiformerV2
39.0
95.5
114.6
306.7
2.94 ×
3.21 ×
7.73
19.77
Appendix
Table S7: Energy-only inference against energy+forces. Energy-only is timed under no autograd graph retained.
Atoms b
LOOP+ (s) a
SCF steps
s per SCF step
18
2.5
35
0.07
38
8.3
39
0.21
66
21.7
38
0.57
102
63.3
45
1.41
146
149.7
43
3.48
Appendix
Table S8: DFT calculations on CPUs: VASP 6.3.2 single points on the same Pt(111)+CO cells, RPBE, cutoff with 450 eV, and 64 MPI ranks on one node.
Figure S11: The wall time consumption in an energy+forces call with same models and hardware as Figure 6 in the manuscript. (a) Share of each call spent building the neighbour list against system size. (b) Throughput against batch size at a fixed 108-atom cell. (c) Cost against the number of edges in the graph at the 5.5A˚ cutoff for two geometries: bulk fcc Pt at 42 edges per atom (filled) and the Pt(111) slab at 33 (open), whose surface atoms are missing a quarter of their neighbours. (d) Ratio of CPU to GPU wall time, per architecture and per size.
Figure S12: Energy-only inference versus energy+forces inference. (a) Time per energy-only call against system size, timed under no autograd graph is built. (b) Forces cost, as the ratio of the two timing modes. (c) Peak GPU memory in both modes. Solid is for the energy+forces mode and open is for the energy-only mode. (d) Throughput in the two modes against each other. The corresponding numbers are in Table S7 .
Architecture
B per
Global B
X(1)
X(16)
Speedup
Efficiency
All-reduce
Mem/GPU
rank
at P =16
(struct./s)
(struct./s)
(%)
(% step)
(GiB)
PaiNN
16
256
305.4
3892.5
12.75
79.7 ± 0.7
19.5
3.22
NequIP
16
256
218.6
2919.7
13.36
83.6 ± 0.2
8.1
1.62
Allegro
16
256
70.3
756.5
10.77
67.3 ± 1.0
41.1
9.24
MACE
16
256
208.8
2869.0
13.74
86.0 ± 0.5
7.5
2.17
EquiformerV2
8
128
30.8
323.2
10.50
65.7 ± 1.6
43.2
24.27
Appendix
Table S9: Training efficiency of different architectures on MPtrj on 1 to 16 NVIDIA A100-SXM4-80GB GPUs. B is each run’s own per-rank batch size, so the global batch is P×B and the work per GPU is fixed; X is the global throughput in structures per second and the efficiency is X(P)/(PX(1)) . The all-reduce column is the part of the gradient reduction that is not hidden behind the backward pass.
X(P) (structures s -1 )
Parallel efficiency (%)
Architecture
B
1
2
4
8
16
2
4
8
16
PaiNN
16
305
577
1106
2042
3892
94.3
90.5
83.5
79.7
NequIP
16
219
431
798
1509
2920
98.5
91.3
86.5
83.6
Allegro
16
70
118
223
410
756
84.4
79.7
73.0
67.3
MACE
16
209
409
775
1470
2869
98.0
92.8
88.1
86.0
EquiformerV2
8
31
55
98
176
323
89.0
79.5
71.3
65.7
Appendix
Table S10: Training global throughput X(P) in structures per second, and parallel efficiency X(P)/(PX(1)) , at every allocation from 1 to 16 A100 GPUs. B is the per-rank batch size, so the global batch at each column is P×B . Every world size times an identical set of structures.
Figure S13: Parallel training on 1–16 A100 GPUs, batch 16 structures per rank, on the cleaned MPtrj database. (a) Speedup X(P)/X(1) against rank count, with the ideal line. (b) Parallel efficiency. (c) Global throughput in structures per second. (d) Where one training step goes at P=16 : data loading and neighbour-list construction, host-to-device transfer, forward and backward, the gradient all-reduce, and gradient clipping plus the optimizer update. The all-reduce share is what separates the curves in (b).
Time per step (s)
Parallel efficiency (%)
N
1 GPU
2 GPUs
4 GPUs
8 GPUs
E2
E4
E8
Y8 (atom step s -1 )
serial (s)
864
0.0896
0.0732
0.0629
0.0854
61
36
13
10,115
0.0151
2,048
0.0990
0.0690
0.0471
0.1240
72
53
10
16,513
0.0330
5,324
0.1905
0.1506
0.1085
0.1236
63
44
19
43,074
0.0789
6,912
0.2286
0.1824
0.1480
0.1327
63
39
22
52,099
0.0994
10,976
0.3307
0.2022
0.1633
0.1488
82
51
28
73,763
0.1587
Appendix
Table S11: The performance of the IANN LAMMPS interface of pair_style iann/multi_gpu on NVIDIA A100-80GB GPUs in Figure 7 in the manuscript is tabulated here. The step time per step on 1 to 8 GPUs, the corresponding efficiencies derived from them, and the Y8 throughput at eight GPUs are listed at each system size. For 8 GPUs, four GPUs sit in one node (NVLink), and eight span two (Slingshot). The last column is the exact serial run of LAMMPS interface with pair_style iann on a single GPU.
Figure S14: PBE-MPtrj: number of structures containing each element. 1,579,091 structures, 49,231,714 atoms, 89 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S15: PBE-MPtrj: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S16: PBE-MPtrj: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0517 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S17: PBE-MPtrj: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0498 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S18: PBE-sAlex: number of structures containing each element. 10,423,284 structures, 107,455,806 atoms, 89 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S19: PBE-sAlex: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S20: PBE-sAlex: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0408 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S21: PBE-sAlex: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0513 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S22: PBE-OMat24: number of structures containing each element. 11,274,650 structures, 157,069,839 atoms, 89 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S23: PBE-OMat24: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S24: PBE-OMat24: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0264 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S25: PBE-OMat24: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.1391 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S26: PBE-MatPES: number of structures containing each element. 434,092 structures, 3,865,819 atoms, 89 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S27: PBE-MatPES: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S28: PBE-MatPES: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0463 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S29: PBE-MatPES: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0930 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S30: PBE-OC22: number of structures containing each element. 8,199,197 structures, 654,340,594 atoms, 57 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S31: PBE-OC22: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S32: PBE-OC22: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0381 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S33: PBE-OC22: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0528 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S34: RPBE-OC20: number of structures containing each element. 1,999,572 structures, 146,467,030 atoms, 56 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S35: RPBE-OC20: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S36: RPBE-OC20: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0263 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S37: RPBE-OC20: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0630 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S38: RPBE-OC25: number of structures containing each element. 7,370,461 structures, 1,064,330,294 atoms, 73 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S39: RPBE-OC25: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S40: RPBE-OC25: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0116 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S41: RPBE-OC25: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0753 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S42: r2SCAN-MPtrj: number of structures containing each element. 238,241 structures, 5,022,951 atoms, 88 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S43: r2SCAN-MPtrj: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S44: r2SCAN-MPtrj: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0206 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S45: r2SCAN-MPtrj: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.0117 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S46: r2SCAN-MatPES: number of structures containing each element. 387,306 structures, 3,040,410 atoms, 89 elements. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S47: r2SCAN-MatPES: total number of atoms of each element. A structure with many atoms of one element counts once in the map above and once per atom here, so the two differ wherever the stoichiometry is skewed. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S48: r2SCAN-MatPES: per-element energy MAE (eV/atom) of the model trained on this database alone, which reaches 0.0446 eV/atom overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
Figure S49: r2SCAN-MatPES: per-element force MAE (eV/Å) of the model trained on this database alone, which reaches 0.1129 eV/Å overall. Color scales are shared across the databases of a functional, so tiles are directly comparable between the maps of this appendix.
School of Materials Science and Engineering & State Key Laboratory of Advanced Refractories & State Key Laboratory of Advanced Special Steels, Shanghai University, Shanghai, 200444, P. R. China.