pyttb.decompositions.cp.gcp.fg

Evaluate Function And Gradient Handles.

pyttb.decompositions.cp.gcp.fg.evaluate(model: ktensor, data: ttb.tensor | ttb.sptensor, weights: np.ndarray | None, function_handle: Literal[None], gradient_handle: function_type) → list[ndarray][source]
pyttb.decompositions.cp.gcp.fg.evaluate(model: ktensor, data: ttb.tensor | ttb.sptensor, weights: np.ndarray | None, function_handle: function_type, gradient_handle: Literal[None]) → float
pyttb.decompositions.cp.gcp.fg.evaluate(model: ktensor, data: ttb.tensor | ttb.sptensor, weights: np.ndarray | None, function_handle: function_type, gradient_handle: function_type) → tuple[float, list[ndarray]]

Evaluate an objective function and/or gradient function.

Parameters:

  • model – Current decomposition.

  • data – Source tensor to decompose.

  • weights – Weighted values for returned tensor. Can be used as a mask.

  • function_handle – Objective function.

  • gradient_handle – Gradient definition.

Returns:

Objective function value and/or gradient function value with respect to model.