Blazing Fast Hyperparameter Tuning

# Initialize the service for training.
# Change the number of instances ‘n_instances’ to match your dataset.
service_obj = CoresetTreeServiceDTC(
    data_params=data_params,
    optimized_for=”training”,
    n_instances=XXX,
    model_cls=XGBClassifier,
)
# Build the Coreset Tree
# The Coreset tree uses the local file system to store its data.
# After this step you will have a new directory .dataheroes_cache
service_obj.build(X=X, y=y)

# To hyperparameter tune, use the library’s built-in grid_search function,
# which would run dramatically faster than GridSearchCV
# Adjust the hyperparameters and scoring function to your needs
# (or use the default LGBMRegressor scoring by setting scoring=None).
param_grid = {
    'learning_rate': [0.1, 0.01],
    'n_estimators': [250, 500, 1000],
    'max_depth': [4, 6]
)

from sklearn.metrics import make_scorer
scoring = make_scorer(mean_squared_error)

optimal_hyperparameters, trained_model = service_obj.grid_search(
    param_grid=param_grid,
    scoring=scoring,
    refit=True,
    verbose=2
)