I am trying to use a pre trained model specifically training on logos . I am using MobileNet for training in logos . If we made our CNN model from scratch then we provide the la