People who code: we want your input. Take the Survey

New answers tagged

0

In order to compare models, you need to define a single metric for the quality of the model. You don't specify but I assume your model is a multiclass classifier: out of 45 labels, for each example, it needs to pick the right one. You can calculate the F1 score of your model for each label vs the rest: how often is it able to correctly predict label 0 vs not ...


0

One option is to plot fewer epochs. There is no useful information after 20 epochs because training and validation performance are the same after that point.


Top 50 recent answers are included