syne_tune.optimizer.schedulers.searchers.multi_fidelity_searcher module

class syne_tune.optimizer.schedulers.searchers.multi_fidelity_searcher.IndependentMultiFidelitySearcher(config_space, searcher='kde', points_to_evaluate=None, random_seed=None, searcher_kwargs=None)[source]

Bases: BaseSearcher

Searcher for the multi-fidelity setting which fits independent models for each resource level as proposed by Falkner et al.

BOHB: Robust and Efficient Hyperparameter Optimization at Scale
S. Falkner and A. Klein and F. Hutter
Proceedings of the 35th International Conference on Machine Learning
Parameters:

  • config_space (Dict[str, Any]) – Configuration space

  • searcher (Union[str, SingleObjectiveBaseSearcher, None]) – Searcher to sample configurations on each rung level.

  • points_to_evaluate (Optional[List[dict]]) – List of configurations to be evaluated initially (in that order).

  • random_seed (Optional[int]) – Seed used to initialize the random number generators.

suggest(**kwargs)[source]

Suggest a new configuration.

Note: Query _next_points_to_evaluate() for initial configs to return first.

Parameters:

kwargs – Extra information may be passed from scheduler to searcher

Return type:

Optional[Dict[str, Any]]

Returns:

New configuration. The searcher may return None if a new configuration cannot be suggested. In this case, the tuning will stop. This happens if searchers never suggest the same config more than once, and all configs in the (finite) search space are exhausted.

on_trial_result(trial_id, config, metric, resource_level)[source]

Updates the model with the latest results of a trial at a specific resource level.

Parameters:

  • trial_id (int) – See on_trial_result()

  • config (Dict[str, Any]) – See on_trial_result()

  • metric (float) – See on_trial_result()

  • resource_level (int) – Resource level where the metric was observed from.

on_trial_complete(trial_id, config, metric, resource_level)[source]

Updates the model with the final results of a completed trial at a specific resource level.

Parameters:

  • trial_id (int) – See on_trial_result()

  • config (Dict[str, Any]) – See on_trial_result()

  • metric (float) – See on_trial_result()

  • resource_level (int) – Resource level where the metric was observed from.