Hi all,
we are using a Parameter Optimization Loop for hyperpararmeter tuning of certain machine learning models. The workflow looks (very roughly) like this:
Question:
Is there any way to obtain (after the loop has terminated) not only the optimized parameters but also the optimized model itself that has been built with these optimal parameters?
Currently I don’t see any other way than having another Learner node after the loop which builds the optimal model again. This seems a bit inefficient to me (because that same model has been created already), especially when the learning part takes long…
don’t think so. You need to use Table Row to Variable to get best parameters and apply them to Learner node.
There is already a ticket in system (Internal reference: AP-11486) for Parameter Optimization Loop End node to save best model and will note there your request.