model.runtime.configs.parameter_config

Parameterization-level runtime configuration.

This module defines how model parameters are provided to the LSM: fixed in config, loaded externally, derived from lookup tables, or generated by NN-based parameterization modules.

Module Contents

Classes

ParameterSourceConfig	Source specification for one model parameter.
NNParameterizationConfig	Configuration for NN-based parameterization modules.
ParameterizationConfig	Top-level parameterization block.

API

class model.runtime.configs.parameter_config.ParameterSourceConfig

Bases: model.runtime.configs.base_config.BaseConfig

Source specification for one model parameter.

source: str

None

value: object

None

bounds: list[float]

‘field(…)’

classmethod from_dict(raw)

validate(name)

class model.runtime.configs.parameter_config.NNParameterizationConfig

Bases: model.runtime.configs.base_config.BaseConfig

Configuration for NN-based parameterization modules.

attri_features: list[str]

‘field(…)’

hidden_dims: list[int]

‘field(…)’

dropout_rate: float

0.2

classmethod from_dict(raw)

validate()

class model.runtime.configs.parameter_config.ParameterizationConfig

Bases: model.runtime.configs.base_config.BaseConfig

Top-level parameterization block.

Holds per-parameter source declarations and the neural network configuration used for nn_global and nn_feature_based parameters.

parameters: dict[str, model.runtime.configs.parameter_config.ParameterSourceConfig]

‘field(…)’

nn: model.runtime.configs.parameter_config.NNParameterizationConfig

‘field(…)’

classmethod from_dict(raw)

validate()