currently the optimization algorithm does as many steps as prescribed.
It would be beneficial to define a convergence threshold to save computational time in cases where only marginal optimization are possible.
i.e. something like:
for ii in range(maximal_gradient_descent_iterations): if convergence_threshold_reached == False: do_opt
currently the optimization algorithm does as many steps as prescribed. It would be beneficial to define a convergence threshold to save computational time in cases where only marginal optimization are possible.
i.e. something like:
for ii in range(maximal_gradient_descent_iterations): if convergence_threshold_reached == False: do_opt