Easier hyper parameter looping for model types?

Is there an easier way to configure the parameter optimisation loop nodes to work with different model types? It appears that you have to figure out each tunable parameter and manually assign them to a flow variable. Since you can’t have the config dialogue for two nodes open at once, it becomes a real pain to have to go back and forth for more complex model types.

If there isn’t an eaiser way already, it would be greate if one could be added. If the parmaeter optimisation loop node could be told what model type its being used on and automatically set up a number of variables, that would be great. It would be even better if the model learner node could then automatically assign the variable values to inputs.

1 Like

Hi @rsherhod ,

Welcome back to the forum!

I understand the challenges you’re facing with configuring parameter optimization loop nodes for different model types. There is a more streamlined way to handle this process.

I recommend using the Parameter Optimization (Table) component, which significantly simplifies hyperparameter tuning. The general workflow involves capturing the model to be optimized using capture nodes along with the tuning parameters. The parameter names must be consistent between the capture and optimization parts of the workflow.

Here is a sample workflow demonstrating how can this be implemented on a Random Forest model.

I hope this helps! If you have any further questions please feel free to ask.


1 Like