Is it possible to show metrics for multiple validation sets during fit_one_cycle()?
For example, in the Adults tutorial, I’d like to see the accuracy of the model on people of each education level. The output would look something like this:
epoch | train_loss | validation_loss | accuracy | accuracy(HS-grad) | accuracy(Bachelors) | accuracy(Some-college) |
---|---|---|---|---|---|---|
0 | 0.366727 | 0.351524 | 0.8 | 0.9 | 0.7 | 0.6 |