hi,
do you think you can add the option to choose different loss functions when modelling? Huber loss function is very effective handling outliers when using decision trees
