Hyperparameter optimization.
Functions
execute_tune(**kwargs) -> Dict[str, Any]
Hyperparameter optimization for any registered algorithm.
Backs the
tuiml_tune tool. Runs grid, random, or Bayesian search over a parameter space, then saves and indexes the best estimator.Parameters
algorithm
str
Registered algorithm class name to tune (arrives via
**kwargs, like all parameters below).
data
str
Dataset to tune on: dataset_id, file path, or built-in name.
method
str
Search strategy:
'grid', 'random', or 'bayesian'.
param_grid
dict
Parameter grid / distributions / space, depending on
method.
cv
int
= 5
Number of cross-validation folds.
scoring
str
= 'accuracy'
Scoring metric name.
n_iter
int
= 10
Number of sampled candidates (random search only).
n_iterations
int
= 50
Number of optimization iterations (Bayesian search only).
random_seed
int
= None
Random seed for reproducible search.
_progress_callback
callable
= None
Internal per-iteration progress hook; stripped from recorded args.
Returns
result
dict
On success:
status ('success'), method, best_params, best_score, cv_results (summary with n_candidates, best_rank and top_5), model_id, model_path, and optionally progress_log. On failure: status ('error'), error and optionally suggestion / error_type.