Hello everyone. I was thinking about using optimization loop nodes on a keras architecture in order to optimize the number of layers and nodes present. Do you think it would be a good idea and, in case, could you please provide me with a workflow where this specific problem is solved with the simplest case, that is Iris.csv datset? Thanks in advance.
Hi @matt0 -
One of our data scientists had this to say about your question:
“Here’s an example where Maarit used the rule engine to twist the parameter optimization loop into working with categorical features. Between that and the normal numeric features you can optimize… I suppose you can optimize an architecture. Although I’m not sure how useful going down that path would be.”
Maybe it would be more useful to focus on different aspects of the network itself, instead of the optimization angle.
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.