Fit

Construct the model with a family and one declaration per family parameter, made with that family’s helper methods. fit_reml() fits the coefficients and estimates the smoothing parameters jointly, by generalised Fellner-Schall updates with optional Newton refinement, and is the normal path; fit() holds the smoothing parameters fixed at the lambdas you pass. diagnose() returns a FitDiagnosticReport that explains how the fit ran and how smoothing stopped: phase timings, iteration and refit counts, and the terminal state of every smoothing component. predictors and family echo the declaration as independent copies. The how-to on fitting a distributional model covers the declaration syntax.

superglm.SuperLSS

Fit several parameters of a response distribution jointly.

fit	Fit coefficients while holding smoothing parameters fixed.
fit_reml	Fit coefficients and estimate smoothing parameters jointly.
diagnose	Explain solver and smoothing behavior from the accepted fitted revision.
predictors	Return independent predictor templates in family parameter order.
family	Return an independent copy of the model's family configuration.
superglm.FitDiagnosticReport	An immutable structured report with compact and full text renderings.