High level API

falcon.AutoML(task: str, train_data: Any, test_data: Optional[Any] = None, features: Optional[Any] = None, target: Optional[Any] = None, manager_configuration: Optional[Union[Dict, str]] = None, config: Optional[Union[Dict, str]] = None) → TaskManager

High level API for one line model training and evaluation.

When calling the following steps will be executed:

  1. task manager object will be initialized;

  2. the model will be trained;

  3. performance summary table is printed (if test set is not provided, random split is done);

  4. the model is saved as an onnx file.

Parameters

  • task (str) – type of the task, currently supported tasks are [tabular_classification, tabular_regression]

  • train_data (Any) – data to be used for training, for tabular classification and regression this can be: path to .csv or .parquet file, pandas dataframe, numpy array, tuple (X,y)

  • test_data (Any, optional) – data to be used for evaluation, for tabular classification and regression this can be: path to .csv or .parquet file, pandas dataframe, numpy array, tuple (X,y)

  • features (Any, optional) – features to be used for training, for tabular classification and regression this can be: list of column names or indexes, by default None

  • target (Any, optional) – target to be used for training, for tabular classification and regression this can be: column name or index, by default None

  • manager_configuration (Union[Dict, str], optional) – task manager configuration to be used (can be used to replace pipeline/learner and/or their arguments), by default None

  • config (Union[Dict, str], optional) – alias for manager_configuration argument

Returns

Task Manager object for the corresponding task.

Return type

TaskManager

falcon.initialize(task: str, data: Any, pipeline: Optional[Type[Pipeline]] = None, pipeline_options: Optional[Dict] = None, extra_pipeline_options: Optional[Dict] = None, features: Optional[Any] = None, target: Optional[Any] = None, **options: Any) → TaskManager

Initializes and returns a task manager object for a given task.

Parameters

  • task (str) – type of the task

  • data (Any) – data to be used for training

  • pipeline (Optional[Type[Pipeline]], optional) – class to be used as pipeline, by default None

  • pipeline_options (Optional[Dict], optional) – arguments to be passed to the pipeline, by default None. These options will overwrite the ones from default_pipeline_options attribute

  • extra_pipeline_options (Optional[Dict], optional) – arguments to be passed to the pipeline, by default None. These options will be passed in addition to the ones from default_pipeline_options attribute. This argument is ignored if pipeline_options is not None

  • features (Any, optional) – features to be used for training, by default None

  • target (Any, optional) – target to be used for training, by default None

Returns

Initialized task manager object

Return type

TaskManager

falcon.run_model(model_path: str, X: ndarray[Any, dtype[ScalarType]]) → Union[List[ndarray[Any, dtype[ScalarType]]], ndarray]

Runs input data through the saved model.

Parameters

  • model_path (str) – model path

  • X (npt.NDArray) – model inputs

Returns

model predictions

Return type

Union[List[npt.NDArray], np.ndarray]