Ablation Paths

This example shows how to use the plugin Ablation Paths. Note that other plugins use the same interfaces and can be used in the same fashion.

from deepcave.plugins.hyperparameter.ablation_paths import AblationPaths
from deepcave.runs.converters.deepcave import DeepCAVERun
from pathlib import Path


if __name__ == "__main__":
    # Instantiate the run
    run = DeepCAVERun.from_path(Path("logs/DeepCAVE/minimal/run_1"))

    objective_id1 = run.get_objective_ids()[0]
    objective_id2 = None  # replace with run.get_objective_ids()[1] for multi-objective importance
    budget_id = run.get_budget_ids()[0]

    # Instantiate the plugin
    plugin = AblationPaths()
    inputs = plugin.generate_inputs(
        objective_id1=objective_id1,
        objective_id2=objective_id2,
        budget_id=budget_id,
        n_hps=100,
        n_trees=100,
        show_confidence=False,
    )
    # Note: Filter variables are not considered.
    outputs = plugin.generate_outputs(run, inputs)

    # Finally, you can load the figure. Here, the filter variables play a role.
    figure1, figure2 = plugin.load_outputs(run, inputs, outputs)
    figure1.write_image("examples/api/ablation_paths_performance.png", scale=2.0)
    if not objective_id2:
        figure2.write_image("examples/api/ablation_paths_improvement.png", scale=2.0)

previous

Importances

next

Record