Skip to content

Release Notes

New Features

  • XGBoost updated to 1.3.1 -> now expected to use native xgboost.dask for models
  • XGBoost compact modeling
  • Sci-kit learn compact modeling and official support

Changes

  • new parameter required when starting training job: “algorithm”; set to either “booster” (default), “xgbclassifier”, or “xgbregressor”
  • predictions and explanations both enabled for XGBoost and Sci-kit learn models