Skip to main content

types.model

ModelServer Objects​

class ModelServer(EnumSpec)

Model server to use for deployment

ModelFramework Objects​

class ModelFramework(EnumSpec)

Framework used to train the model

TrainingContext Objects​

class TrainingContext()

Context of the model request.

features_and_labels​

def features_and_labels() -> pd.DataFrame

Get the features and labels for the model

ModelSpec Objects​

class ModelSpec(RaptorSpec)

Specification of a model

Arguments:

  • keys: List of keys to use for training
  • freshness: How fresh the data should be
  • staleness: How stale the data can be
  • timeout: How long to wait for data
  • features: List of features to use for training
  • label_features: List of label features to use for training
  • key_feature: Feature to use as key
  • model_framework: Framework used to train the model
  • model_server: Model server to use for deployment
  • training_function: Function to use for training
  • exporter: Exporter to use for exporting the model
  • model_framework_version: Version of the model framework
  • runtime: Runtime to use for training