Example of fitting a ModelingToolkit model to data

A common use-case for ModelingToolkit is to fit the model to data, thereby inferring the values of parameter and initial conditions. Often, one would like to change whether a parameter or initial condition is allowed to be adjusted in the optimization process. This is an example of how to accomplish this task. It also helps work around the slowness of access to solutions via the sol[x] approach, which makes naive loss function evaluations too slow for optimization use.