Skip to main content

limen.sfd.reference_architecture

Model-function implementations and templates that own the training logic behind each SFD.

Canonical docs

What this package owns

Owns the reference model classes (and their function forms) that implement fit/predict for each decoder family, without the packaged data pipeline. Does not own the experiment shape or manifest (owned by the matching limen.sfd.foundational_sfd module) or experiment execution.

Key entry points

Entry pointUse caseNotes
ReferenceModelBase class for reference modelsShared fit/predict contract
LogRegBinary / logreg_binaryLogistic-regression binary modelClass and function forms
DLinearRegressor / dlinear_regressorLinear time-series regressorClass and function forms
LightGBMBinary / lightgbm_binaryLightGBM binary classifierClass and function forms
XGBoostRegressor / xgboost_regressorXGBoost regressorClass and function forms
RandomBinary / random_binaryBaseline random classifierClass and function forms
RuleBasedStrategy / rule_basedPredicate-driven, non-learned strategyUses limen.sfd.rule_based
TabPFNBinary / tabpfn_binaryTabPFN binary classifierOptional; requires tabpfn

Adjacent modules

  • limen.sfd.foundational_sfd pairs each of these models with an experiment shape and manifest.
  • limen.inference.Sensor wraps a ReferenceModel for live bar-by-bar prediction.
  • limen.metrics is commonly used inside these model functions to score outputs.

Quick orientation

reference_architecture/
├── base.py # ReferenceModel base class
├── logreg_binary.py
├── dlinear_regressor.py
├── lightgbm_binary.py
├── xgboost_regressor.py
├── random_binary.py
├── rule_based.py
└── tabpfn_binary.py # Optional

Things to know

  • Each module exposes both a class (e.g. LogRegBinary) and a function form (e.g. logreg_binary); pick the class for custom subclassing and the function for manifest wiring.
  • These models are the template for custom SFD work: subclass ReferenceModel or copy a module to start.
  • A reference architecture owns the training logic; the matching foundational SFD owns the manifest and data pipeline.