Feature Request: Give option to limit the training time or number of iterations

[JackyKoh]

Could we get an option to limit the amount of training done by time or iterations with the auto_ml algorithm? Currently there is the option to stop "GradientBoosting" with keyboard interrupt, could we get instead or also an option/function argument from the library to do it by a timeout or max_num_iters argument as well and maybe for more algorithms as well.

Thank you 👍