Smooth Imitation Learning via Smooth Costs and Smooth Policies