Metric exp_rmspe vs. RMSPE?

In lesson 6 Rossman, we have this TabularLearner created:
learn = tabular_learner(data, layers=[1000,500], ps=[0.001,0.01], emb_drop=0.04, y_range=y_range, metrics=exp_rmspe)

Can any one explain what the ‘exp’ in exp_rmspe does, so I can know what’s the difference between it and rmspe?


preds = torch.exp(preds) and same for targets in exp_rmspe. Rest is same.

For larger differences we get larger loss values.

thanks Kushaj

1 Like