Hi, I am fitting a LinearRegression with MLForecast applying some target_tranformations = [Log1p, LocalStandardScaler()]. Everything works fine, so no issues with the code. My problem is that I would like to have the predictions in the transformed scale, hence with no inverse transformations applied. Is there a way to do that without manually applying the transformations? Thanks

Hey. The easiest way is to apply the transformations yourself. If you really want to you can temporarily remove the target transforms attribute, that should make it so they're not restored when predicting:

targ_tfms = mlf.ts.target_transforms
mlf.ts.target_transforms = None
raw_preds = mlf.predict(...)
# restore the target_transforms attribute if you want

Thank you so much for your answer!