automl_leaderboard()
: Returns the AutoML Leaderboard from the AutoML run when the model was created.automl_update_model()
: Makes it easy to swap out models from the AutoML Leaderboard.h2o.init()
to skip on CRAN per H2O (Erin LeDell) suggestion. The original issue was related to libxgboost4j_gpu*.so
being left on the user system when the user’s temp library.This is the initial release with key functions:
automl_reg()
: Uses H2O AutoML as a backend for forecasting.save_h2o_model()
and load_h2o_model()
: Saving and loading H2O models and Workflow/Parsnip objects