MLPotential Keywords¶
Engine MLPotential¶
AIMNet2- Type:
Block
- Recurring:
False
- Description:
Options for the AIMNet2 machine learning potential backend.
CompileModel- Type:
Bool
- Default value:
No
- Description:
Compile the AIMNet2 model. This can improve performance for repeated calculations (e.g. during molecular dynamics), but increases startup cost.
EnsembleSize- Type:
Integer
- Default value:
4
- Value Range:
2 <= value <= 4
- Description:
Number of members in the ensemble. Uncertainty estimates improve with the number of members, but computational cost scales linearly.
LongRange- Type:
Block
- Recurring:
False
- Description:
Long-range electrostatics options.
DSF- Type:
Block
- Recurring:
False
- Description:
Damped shifted force Coulomb settings.
Alpha- Type:
Float
- Default value:
0.2
- Value Range:
value > 0
- Description:
Damping parameter for the DSF Coulomb method.
Cutoff- Type:
Float
- Default value:
15.0
- Unit:
Angstrom
- Value Range:
value > 0
- Description:
Cutoff distance in Angstrom for the DSF Coulomb neighbor list.
Ewald- Type:
Block
- Recurring:
False
- Description:
Accuracy settings for Ewald and PME Coulomb summation.
Accuracy- Type:
Float
- Default value:
1e-06
- Value Range:
value > 0
- Description:
Target accuracy for Ewald and PME summation. Smaller values are more accurate and more expensive.
Type- Type:
Multiple Choice
- Default value:
Auto
- Options:
[Auto, Simple, DSF, Ewald, PME]
- Description:
Long-range Coulomb method. Auto uses the AIMNet2 model default (typically simple for aperiodic systems, DSF for periodic systems). Simple uses all-pairs pairwise Coulomb, and is suitable for small aperiodic systems. DSF (damped shifted force) Coulomb with a finite cutoff is an efficient method for periodic and large aperiodic systems. Ewald and PME (Particle mesh Ewald) is for high accuracy periodic electrostatics.
UseEnsemble- Type:
Bool
- Default value:
No
- Description:
Whether to use an ensemble of independently trained model members for calculations. This gives uncertainty estimates for energies and forces, but at increased computational cost.
Backend- Type:
Multiple Choice
- Options:
[AIMNet2, FAIRChem, M3GNet, MACE, MatGL, NEP, NequIP, TorchANI]
- Description:
The machine learning potential backend.
Device- Type:
Multiple Choice
- Default value:
- Options:
[, cpu, cuda:0, cuda:1]
- Description:
Device on which to run the calculation (e.g. cpu, cuda:0).
If empty, the device can be controlled using environment variables for TensorFlow or PyTorch.
FAIRChem- Type:
Block
- Recurring:
False
- Description:
Options for the FAIRChem machine learning potential backend.
ModelTask- Type:
String
- Default value:
- Description:
Model task to use if a custom UMA/eSEN model is supplied via a parameter file (e.g. ‘OC20’, ‘OC22’, ‘OC25’, ‘ODAC’, ‘OMat’, ‘OMC’, ‘OMol’). Ignored if a specific FAIRChem model is selected.
MACE- Type:
Block
- Recurring:
False
- Description:
Options for the MACE machine learning potential backend.
DataType- Type:
Multiple Choice
- Default value:
float32
- Options:
[float32, float64]
- Description:
Using
float32is faster but less accurate, and generally recommended for MD. Conversely usingfloat64is slower but more accurate, and recommended for geometry optimization.
EnableCuEquivariance- Type:
Bool
- Default value:
Yes
- Description:
Enable CUDA-accelerated cuEquivariance library for equivariant neural networks, if CUDA available.
ModelHead- Type:
String
- Default value:
- Description:
Model head to use if a custom MACE model is supplied via a parameter file (e.g. ‘omat_pbe’, ‘omol’, ‘spice_wB97M’, ‘ωB97M-D3(BJ)’, ‘rgd1_b3lyp’, ‘oc20_usemppbe’, ‘matpes_r2scan’). Ignored if a specific MACE model is selected.
MLDistanceUnit- Type:
Multiple Choice
- Default value:
Auto
- Options:
[Auto, angstrom, bohr]
- GUI name:
Internal distance unit
- Description:
Unit of distances expected by the ML backend (not the ASE calculator). The ASE calculator may require this information.
MLEnergyUnit- Type:
Multiple Choice
- Default value:
Auto
- Options:
[Auto, Hartree, eV, kcal/mol, kJ/mol]
- GUI name:
Internal energy unit
- Description:
Unit of energy output by the ML backend (not the unit output by the ASE calculator). The ASE calculator may require this information.
Model- Type:
Multiple Choice
- Default value:
ANI-2x
- Options:
[Custom, AIMNet2-B973c, AIMNet2-NSE, AIMNet2-Pd, AIMNet2-Rxn, AIMNet2-wB97MD3, ANI-1ccx, ANI-1x, ANI-2x, eSEN-S-Con-OMol, M3GNet-UP-2022, MACE-MP-0-Large, MACE-MP-0-Medium, MACE-MP-0-Small, MACE-MPA-0, NEP89, QET-PBE-2025, QET-r2SCAN-2025, TensorNet-PBE-M-2025, TensorNet-r2SCAN-M-2025, UMA-S-1.2-OC20, UMA-S-1.2-OC22, UMA-S-1.2-OC25, UMA-S-1.2-ODAC, UMA-S-1.2-OMat, UMA-S-1.2-OMC, UMA-S-1.2-OMol]
- Description:
Select a pre-parameterized or custom model.
AIMNet2 variants: best for fast calculations of small, drug-like molecules; limited to systems of 14 elements (H, B, C, N, O, F, Si, P, S, Cl, As/Pd, Se, Br, I).
ANI-(1x/1ccx/2x): best for very fast calculations of organic molecules; limited to elements H, C, N, O (ANI-1x/1ccx), F, S, Cl (ANI-2x).
eSEN-S-Con-OMol: best for highly accurate calculations of diverse organic and bio-relevant molecules; not intended for calculations on periodic inorganic bulk materials.
M3GNet-UP-2022: best for fast calculations of inorganic crystalline materials; not designed for accurately modeling small organic molecules or biomolecules.
MACE-MP-0-(Small/Medium/Large): best for accurate periodic calculations of inorganic materials; size trades speed/accuracy; not designed for accurately modeling small organic molecules or biomolecules. MACE-MPA-0 has improved accuracy vs MP-0.
NEP89: Neuroevolution potential for neutral systems.
UMA-S-1.2 variants: best for high accuracy calculations on a broad range of systems; choose from OC20 (adsorption and surface chemistry), OC22 (oxide catalysis), OC25 (electrocatalysis), ODAC (adsorption in porous frameworks), OMat (inorganic materials), OMC (organic molecular crystals), OMol (molecules, biomolecules, metal complexes, electrolytes); can be computationally expensive compared to other, more targeted models.
Set Custom to choose a backend and provide your own parameters.
NumThreads- Type:
String
- Default value:
- GUI name:
Number of threads
- Description:
Number of threads.
If not empty, OMP_NUM_THREADS will be set to this number; for PyTorch-engines, torch.set_num_threads() will be called.
ParameterDir- Type:
String
- Default value:
- GUI name:
Parameter directory
- Description:
Path to a set of parameters for the backend, if it expects to read from a directory.
ParameterFile- Type:
String
- Default value:
- Description:
Path to a set of parameters for the backend, if it expects to read from a file.
UnpairedElectrons- Type:
Integer
- Default value:
0
- Value Range:
value >= 0
- GUI name:
Spin polarization
- Description:
The number of unpaired electrons in the system for a spin unrestricted calculation. The spin multiplicity is taken as this value plus one.
Unrestricted- Type:
Bool
- Default value:
No
- Description:
Enables spin unrestricted calculations, passing spin information to the machine learning model. Only applicable to ‘AIMNet2-NSE’, ‘UMA-S-1.2-OMol’, ‘eSEN-S-Con-OMol’ and custom FAIRChem models.