Curated `mlmm all` Starter Snapshot¶

This page is a curated, non-exhaustive starter snapshot for mlmm all. It shows a common subset of keys whose values are pinned to (and equal) their runtime owners; it is not the full configuration schema. For every configurable section and option, see the YAML Reference.

Source template: .github/scripts/generate_reference.py::_ALL_TEMPLATE
Template digest: e777ecfc5686

Included Sections¶

Section
`calc`
`freq`
`thermo`
`dft`

Starter Template¶

# Starter config for `mlmm all`

calc:
  backend: uma              # ML backend: uma, orb, mace, aimnet2
  orb_model: orb_v3_conservative_omol  # ORB model name (when backend=orb)
  orb_precision: float64    # ORB precision default (when backend=orb; "float32-high" = TF32 matmul, also via --precision fp32; legacy "float32" alias accepted)
  mace_model: MACE-OMOL-0   # MACE model path or name (when backend=mace)
  mace_dtype: float64       # MACE dtype, e.g. float32 / float64 (when backend=mace)
  aimnet2_model: aimnet2    # AIMNet2 model name (when backend=aimnet2)

freq:
  max_write: 10
  amplitude_ang: 0.8
  n_frames: 20
  sort: value

thermo:
  temperature: 298.15
  pressure_atm: 1.0
  symmetry_number: null

dft:
  func_basis: wb97m-v/def2-tzvpd
  max_cycle: 100
  conv_tol: 1.0e-9
  grid_level: 3

Scalar Defaults¶

Each scalar is pinned to (and equals) the runtime owner shown.

Key	Type	Default	Runtime owner
`calc.backend`	`str`	`'uma'`	`MLMM_CALC_KW["backend"]`
`calc.orb_model`	`str`	`'orb_v3_conservative_omol'`	`MLMM_CALC_KW["orb_model"]`
`calc.orb_precision`	`str`	`'float64'`	`MLMM_CALC_KW["orb_precision"]`
`calc.mace_model`	`str`	`'MACE-OMOL-0'`	`MLMM_CALC_KW["mace_model"]`
`calc.mace_dtype`	`str`	`'float64'`	`MLMM_CALC_KW["mace_dtype"]`
`calc.aimnet2_model`	`str`	`'aimnet2'`	`MLMM_CALC_KW["aimnet2_model"]`
`freq.max_write`	`int`	`10`	`FREQ_KW["max_write"]`
`freq.amplitude_ang`	`float`	`0.8`	`FREQ_KW["amplitude_ang"]`
`freq.n_frames`	`int`	`20`	`FREQ_KW["n_frames"]`
`freq.sort`	`str`	`'value'`	`FREQ_KW["sort"]`
`thermo.temperature`	`float`	`298.15`	`THERMO_KW["temperature"]`
`thermo.pressure_atm`	`float`	`1.0`	`THERMO_KW["pressure_atm"]`
`thermo.symmetry_number`	`NoneType`	`None`	`THERMO_KW["symmetry_number"]`
`dft.func_basis`	`str`	`'wb97m-v/def2-tzvpd'`	`DFT_KW["func_basis"]`
`dft.max_cycle`	`int`	`100`	`DFT_KW["max_cycle"]`
`dft.conv_tol`	`float`	`1e-09`	`DFT_KW["conv_tol"]`
`dft.grid_level`	`int`	`3`	`DFT_KW["grid_level"]`

Scan Spec Shapes¶

Accepted by scan, scan2d, and scan3d with -s/--scan-lists.

# scan (1D staged)
one_based: false
stages:
  - - [1, 2, 1.65]
  - - [2, 3, 2.30]

# scan2d
one_based: false
pairs:
  - [1, 2, 1.40, 2.20]
  - [2, 3, 1.20, 2.00]

# scan3d
one_based: false
pairs:
  - [1, 2, 1.40, 2.20]
  - [2, 3, 1.20, 2.00]
  - [3, 4, 1.00, 1.80]