quantlab.model.library_model

Library variant of the model layer: LibraryModel and its rows.

LibraryModel is the base class of heads whose library trains itself (XGBoost, pytabkit): there is no epoch loop, the head’s _fit_model receives flat Rows (one per (bar, symbol) cell with a valid training target) and uses the library’s native early stopping. Checkpoints are .joblib files written with joblib. Shipped heads live in quantlab/model/predefined.

Classes

LibraryModel	Numpy variant for tree models and other non-torch libraries.
Rows	The training or validation rows of one fit.