Fine Tunning Speaker Identification

Anyone had any success trying to fine tune pre-trained speaker ID modules, i’m kind of lost. Any direction at all would be really helpfull.