DeepchecksModelVersionClient

class DeepchecksModelVersionClient

Client to interact with a given model version, including all functions to send data.

Parameters

hoststr

The deepchecks monitoring API host.

model_version_idint

The id of the model version.

__init__(model_version_id: int, model: Dict[str, Any], api: API)

__new__(*args, **kwargs)

Attributes

Methods

DeepchecksModelVersionClient.get_feature_importance()	Get the feature importance as a pandas Series.
DeepchecksModelVersionClient.get_production_data(...)	Get DataFrame or Deepchecks dataset and predictions for a model version production data on a specific window.
DeepchecksModelVersionClient.get_reference_data([...])	Get DataFrame or Deepchecks dataset and predictions for a model version reference data.
DeepchecksModelVersionClient.log_batch(...)	Log batch of samples.
DeepchecksModelVersionClient.log_sample(...)	Add a data sample for the model version update queue.
DeepchecksModelVersionClient.send([...])	Send all the aggregated samples for upload or update.
DeepchecksModelVersionClient.set_feature_importance(...)	Set model version feature importance.
DeepchecksModelVersionClient.time_window_statistics([...])	Get statistics on uploaded samples for the model version in a provided time window.
DeepchecksModelVersionClient.upload_reference(...)	Upload reference data.
DeepchecksModelVersionClient.validate([...])	Used in create version when the version already exists to verify it's matching the user passed properties.

previous

DeepchecksModelClient.version

next

DeepchecksModelVersionClient.get_feature_importance