openml.extensions.keras.KerasExtension

class openml.extensions.keras.KerasExtension

Connect Keras to OpenML-Python.

classmethod can_handle_flow(flow: 'OpenMLFlow') → bool

Check whether a given flow describes a Keras neural network.

This is done by parsing the external_version field.

Parameters

flowOpenMLFlow

Returns

bool

classmethod can_handle_model(model: Any) → bool

Check whether a model is an instance of keras.models.Model.

Parameters

modelAny

Returns

bool

compile_additional_information(self, task: 'OpenMLTask', additional_information: List[Tuple[int, int, Any]]) → Dict[str, Tuple[str, str]]

Compiles additional information provided by the extension during the runs into a final set of files.

Parameters

taskOpenMLTask

The task the model was run on.

additional_information: List[Tuple[int, int, Any]]

A list of (fold, repetition, additional information) tuples obtained during training.

Returns

filesDict[str, Tuple[str, str]]

A dictionary of files with their file name and contents.

create_setup_string(self, model: Any) → str

Create a string which can be used to reinstantiate the given model.

Parameters

modelAny

Returns

str

flow_to_model(self, flow: 'OpenMLFlow', initialize_with_defaults: bool = False) → Any

Initializes a Keras model based on a flow.

Parameters

flowmixed

the object to deserialize (can be flow object, or any serialized parameter value that is accepted by)

initialize_with_defaultsbool, optional (default=False)

If this flag is set, the hyperparameter values of flows will be ignored and a flow with its defaults is returned.

Returns

mixed

get_version_information(self) → List[str]

List versions of libraries required by the flow.

Libraries listed are Python, keras, numpy and scipy.

Returns

List

instantiate_model_from_hpo_class(self, model: Any, trace_iteration: openml.runs.trace.OpenMLTraceIteration) → Any

Instantiate a base_estimator which can be searched over by the hyperparameter optimization model (UNUSED)

Parameters

modelAny

A hyperparameter optimization model which defines the model to be instantiated.

trace_iterationOpenMLTraceIteration

Describing the hyperparameter settings to instantiate.

Returns

Any

is_estimator(self, model: Any) → bool

Check whether the given model is a Keras neural network.

This function is only required for backwards compatibility and will be removed in the near future.

Parameters

modelAny

Returns

bool

model_to_flow(self, model: Any) → 'OpenMLFlow'

Transform a Keras model to a flow for uploading it to OpenML.

Parameters

modelAny

Returns

OpenMLFlow

obtain_parameter_values(self, flow: 'OpenMLFlow', model: Any = None) → List[Dict[str, Any]]

Extracts all parameter settings required for the flow from the model.

If no explicit model is provided, the parameters will be extracted from flow.model instead.

Parameters

flowOpenMLFlow

OpenMLFlow object (containing flow ids, i.e., it has to be downloaded from the server)

model: Any, optional (default=None)

The model from which to obtain the parameter values. Must match the flow signature. If None, use the model specified in OpenMLFlow.model.

Returns

list

A list of dicts, where each dict has the following entries: - oml:name : str: The OpenML parameter name - oml:value : mixed: A representation of the parameter value - oml:component : int: flow id to which the parameter belongs

seed_model(self, model: Any, seed: Union[int, NoneType] = None) → Any

Not applied for Keras, since there are no random states in Keras.

Parameters

modelkeras model

The model to be seeded

seedint

The seed to initialize the RandomState with. Unseeded subcomponents will be seeded with a random number from the RandomState.

Returns

Any