just shows you how little control of output these labs actually have. They are training two models that kind of ended being the same so whatever they were doing specifically didnt make much difference.