Member Junction
    Preparing search index...

    Interface MLModelInferenceProcessorOptions

    Construction options for MLModelInferenceProcessor.

    interface MLModelInferenceProcessorOptions {
        autoTrain?: MLAutoTrainConfig;
        autoTrainer?: IModelAutoTrainer;
        datedSources?: DatedSourceSpec[];
        deps: MLInferenceDeps;
        modelId: string;
        primaryKeyField?: string;
    }
    Index

    Properties

    autoTrain? autoTrainer? datedSources? deps modelId primaryKeyField?

    Properties

    autoTrain?: MLAutoTrainConfig

    Opt-in train-if-untrained behavior (see MLAutoTrainConfig). When set together with autoTrainer, a run whose model needs training trains the declared pipeline, publishes the result, repoints the run's Record Process and bindings at it, and then scores — instead of refusing the run. When absent, a model that needs training refuses the run with a clear message.

    autoTrainer?: IModelAutoTrainer

    The trainer that fulfils autoTrain. Required for auto-train to take effect.

    datedSources?: DatedSourceSpec[]

    Optional dated sources supplying point-in-time ("as-of") features, when the model's pipeline uses them. Carried through to the FeatureAssembly executor.

    deps: MLInferenceDeps

    The injected seam bundle (model loader, artifact loader, sidecar).

    modelId: string

    The MJ: ML Models id this processor scores with.

    primaryKeyField?: string

    Primary-key field on the target entity (defaults to ID).

    Settings

    Member Visibility

    On This Page

    Properties