Problem

class t3toolbox.backend.optimizers.Problem

A fixed-rank least-squares fitting problem min_x ½‖S(x) - data‖² for a sampling op S on geometry geom. Layout-agnostic: it holds the operator (kind), the geometry, and the FULL (sample, data). local_model / objective linearize / evaluate at a point on the full data, or on an explicitly-passed minibatch (sample, data) (e.g. from a draw). The Problem itself owns no minibatch-layout logic – where the sample stack W lives, how to slice it – that is the kind’s (kind.take for the default flat draw) or the user’s draw.

geom: GeometryOps

kind: t3toolbox.backend.common.typ.Any

sample: t3toolbox.backend.common.typ.Any

data: t3toolbox.backend.common.typ.Any

regularizer: t3toolbox.backend.common.typ.Any = None

Methods

local_model(x_cores[, sample, data])	Linearize at x_cores on the full data (sample=None) or an explicit minibatch.
objective(x_cores[, sample, data])	½‖S(x)-data‖² (+ ρ(x) if regularized) on the full data (sample=None) or an explicit