Urban traffic is a complex process that is often described by macroscopic flow models. Anyway, the parameters identification of these models remains a heavy work. This paper proposes neural networks architectures that are inspired from the general form of the well-known traffic model but which...