Is there any example of a continuous differentiable function out there, in which the gradient method with Armijo's stepsize-rule doesn't converge?
I found it pretty hard to create one myself because the Armijo-Rule isn't that easy to handle with.
I would be glad for some help, Thanks!