Skip to content
Last updated

AutoML Workflow Operators

AutoML adds workflow operators (base & solution packages). For general operator concepts, see Treasure Workflow Reference.

Base Package Operators

  • ml_train>
    Train: takes a TD table as input, builds multiple models, ensembles them, stores model artifacts internally (S3) for inference.
  • ml_predict>
    Predict: takes test data + model to produce predictions, leaderboards, and statistics as a TD table.

Currently these apply to AutoGluon notebooks (gluon_train, gluon_predict).

Custom Solution Package Operator

  • ipynb>
    General Python notebook task; executes solution‑specific logic defined by chosen package.