TabularData RMSE reported?

Hi, I am using a tabular_learner in order to try and predict (insurance) exposure for a given set of vehicle attributes and was wondering what the reported error is?

After calling lr_find() and then fitting the model using this selected learning rate I see a train_loss, valid_loss and root_mean_squared_error. The first 2 are pretty self-explanatory, however, I am not sure what the last metric is calculated on. Help, please. Thanks!

Thanks. I know what RMSE is, I was wondering what this is being evaluated on? The entire dataset? Seems pretty low…

It’s being evaluated on your validation set.

1 Like