Can Hyperas natively return data regarding all models ran during an optimization cycle?

[michaelmontalbano]

For example, if I test dropout and l2_regularization through hyperas optimization, is there a way to return the set of data which relates the loss metrics of each NN configuration? Or would I have to hardcode that utility myself, likely using pandas?

[JonnoFTW]

Everything returned in the function you want to minimise should be accessible from the Trials object. Of course, you can always have your function output to disk or a database or w/e you like.