Network ODE Model; Type error with ForwardDiff Dual

The initial condition must be a dual number in order to push forward the derivatives with respect to parameters on the state. But if you differentiate by the parameters, then p is Dual while u0 isn’t. So we have a consistency check to upconvert u0 in this case. We missed the specific case though where p is a scalar (always assumed it was an abstractarray, tuple, or named tuple).