My question is whether the returned training loss values in R package xgboost version 1.2.0.1 contain regularization terms. Specifically, in the following example, does trai