mattertune.configs.normalization

class mattertune.configs.normalization.MeanStdNormalizerConfig(*, only_for_target=True, mean, std)[source]
Parameters:

  • only_for_target (bool)

  • mean (float)

  • std (float)

only_for_target: bool

Whether the normalizer should only be applied to the target property or to both predictions and targets.

mean: float

The mean of the property values.

std: float

The standard deviation of the property values.

create_normalizer_module()[source]
Return type:

MeanStdNormalizerModule

class mattertune.configs.normalization.NormalizerConfigBase(*, only_for_target)[source]
Parameters:

only_for_target (bool)

only_for_target: bool

Whether the normalizer should only be applied to the target property or to both predictions and targets.

abstract create_normalizer_module()[source]
Return type:

NormalizerModule

class mattertune.configs.normalization.PerAtomNormalizerConfig(*, only_for_target=False)[source]
Parameters:

only_for_target (bool)

only_for_target: bool

Whether the normalizer should only be applied to the target property or to both predictions and targets.

create_normalizer_module()[source]
Return type:

NormalizerModule

class mattertune.configs.normalization.PerAtomReferencingNormalizerConfig(*, only_for_target=True, per_atom_references)[source]
Parameters:

  • only_for_target (bool)

  • per_atom_references (Mapping[int, float] | Sequence[float] | Path)

only_for_target: bool

Whether the normalizer should only be applied to the target property or to both predictions and targets.

per_atom_references: Mapping[int, float] | Sequence[float] | Path

The reference values for each element.

  • If a dictionary is provided, it maps atomic numbers to reference values

  • If a list is provided, it’s a list of reference values indexed by atomic number

  • If a path is provided, it should point to a JSON file containing the references

create_normalizer_module()[source]
Return type:

NormalizerModule

class mattertune.configs.normalization.RMSNormalizerConfig(*, only_for_target=True, rms)[source]
Parameters:

  • only_for_target (bool)

  • rms (float)

only_for_target: bool

Whether the normalizer should only be applied to the target property or to both predictions and targets.

rms: float

The root mean square of the property values.

create_normalizer_module()[source]
Return type:

RMSNormalizerModule