probe_kind

t3toolbox.backend.fitting.probe_kind(weight=None)

def probe_kind(
        weight: typ.Optional[typ.Any] = None,  # per-mode residual weight ω, (d,) / (d,1); None = 1 (unweighted)
) -> SamplingKind:                             # the vector-valued `probe` kind (optionally per-mode weighted)

The probe sampling kind (vector-valued: one free mode per probe), optionally per-mode weighted. Mode weighting is the order-0 special case of the same residual-weight machinery as the derivative kinds: the objective is ½ Σ_i ‖ω_i z_i‖² over the d per-mode probe residuals z_i, so ω (a per-mode scalar) enters sumsq (×ω) and transpose (×ω²) only. weight=None is the plain unweighted probe (PROBE). Plain probe has no order axis, so the weight is a 1-D (d,) per-mode vector – the frontend enforces that (rejecting a 2-D (d, 1); see t3toolbox.fitting.probe_model()).

Parameters:

weight (t3toolbox.backend.common.typ.Optional[t3toolbox.backend.common.typ.Any])

Return type:

SamplingKind