When doing early-stopping, should i consider improvement of validation loss or validation accuracy ?
If i have two models : Model A : Validation Accuracy = 95% Validation Loss = 0.20 Model B : Validation Accuracy = 96% Validation Loss = 0.21 Which model is better? These values are just made up but im doing some tests and normally i have to compare this kind of models, Model A has better Loss but worse accuracy. Then which model is better? -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
