syndi.task_evaluator.Task_Evaluator.evaluate_task

Task_Evaluator.evaluate_task(metrics=None)[source]

Run benchmark testing on a task. Save intermedia data, trained models, and optimized hyperparameters. Return testing results.

Parameters:
  • task (Task) – a task instance storing meta information of the task.

  • metrics (list) – a list of strings to identify the metric functions.

  • output_path (str) – a directory path to store the intermedia data, model and hyperparametes.

  • agnostic_metrics (boolean) – whether to record dataset agnostic metrics in results

Returns:

benchmarking results of each run.

Return type:

list