Skip to main content
def AgentPredictorWorker.fit(self, inputs, fx_choice=REG.AIG) [source]
Freeze and package the champion agent for production. Resolves the agent, binds encoders, re-fits the model on the curated feature subset, and saves the deployment artifact. After this call, the object is immutable and ready for transform / preprocess.

​
Parameters

inputs : AgentPredictorInput
Training/test data, feature subset, and agent choice.
 fx_choice : str
Fitter choice key.

​
Returns

AgentPredictor
Fitted, deployment-ready champion agent.

​
Raises

  • ValueError: If train_set is empty or features is empty.