Properties

• experiment_id UUID of the original Experiment that generated this Model

• datasets List of (ModelDataSchema) with ids, names, and repartition of DatasetVersion used by original Experiment

• parameters Parameters used by original Experiment to generated Model

Methods

get_infos

get_infos()

Description

Retrieve some infos about this context

Examples

infos = model_context.get_infos()

Returns

A dict with experiment, datasets and parameters

retrieve_experiment

retrieve_experiment()

Description

get_experiment

get_experiment()

Description

Retrieve source experiment of this context

It will raise an exception if this context has no experiment source

Examples

exp = model_context.get_experiment()

Returns

An Experiment object

get_dataset_version

get_dataset_version(
   name: str
)

Description

Retrieve dataset version used to train or evaluate the model, by the name given when attached experiment and dataset.

It will raise an exception if this context has no dataset attached with given name.

Examples

dataset_version = model_context.get_dataset_version("train")

Arguments

• name (str) : Name of the dataset version attached to the experiment

Returns

A DatasetVersion object

retrieve_datasets

retrieve_datasets()

Description

Retrieve datasets used to train and evaluate (or else) the model

It will raise an exception if this context has no data

Examples

dataset_versions = model_context.retrieve_datasets()

Returns

A dict with dataset name as key and DatasetVersion as value