[migrated from https://github.com/dmlc/xgboost/issues/3457]
I noticed that in XGBoost4j-Spark source code, the train method of the XGBoostClassifier class doesn’t have the watchlist features. While in XGBoost4J, the watchlist features is implemented. Do you have plans to add it to the XGBoost4j-Spark package? Also, any plans to add cross validation to XGBoost4j-Spark?