The AmplModel, SlackModel and SlackModelNn have not changed.
I have just added the models of Maxime, that were working in the separate configuration.
I propose to merge the repositories before we do the adjustments, so that each contributor can work on their part and discuss on how to simplify the classes.
The NlpModel has a new function fgobj to compute more efficiently the objective and its gradient when they are required together by the solver.
The functions of NlpModel like gconprod and those concerning the scaling are not implemented in Maxime's model.
The AmplModel, SlackModel and SlackModelNn have not changed. I have just added the models of Maxime, that were working in the separate configuration. I propose to merge the repositories before we do the adjustments, so that each contributor can work on their part and discuss on how to simplify the classes. The NlpModel has a new function fgobj to compute more efficiently the objective and its gradient when they are required together by the solver. The functions of NlpModel like gconprod and those concerning the scaling are not implemented in Maxime's model.