I am not denying what you are saying but I am interested in what models you were using for these tasks?