post_processor

class MethodType(value)

Bases: str, Enum

An enumeration.

ASGD = 'asgd'

Adam = 'adam'

BFGS = 'bfgs'

GradDescent = 'grad-descent'

LBFGS = 'lbfgs'

class PostProcessor

Bases: ABC

A Hypothetical PostProcessor Class Interface.

func_post(c, *args)

Generates the objective function as a numpy scalar.

Parameters:

• c (torch.tensor) – The values for each

• solver. (variable of the problem in the solution found by the) –

Returns:

Objective function.

Return type:

torch.tensor

func_post_jac(c, *args)

Calculates the Jacobian of the objective function as a numpy vector for the post-processing if the post processing is performed using the Jacobian. The post-processing can still be performed without the Jacobian but having it for some post-processing methods can improve the performance of the post-processing. Jacobian can only be used with the numpy post-processing methods of “BFGS”.

Parameters:

• c (torch.tensor) – The values for each

• solver. (variable of the problem in the solution found by the) –

Returns:

Objective function.

Return type:

torch.tensor

abstract postprocess()

An abstract interface method.