Add init_model as a fit parameter

optgbm/sklearn.py

@@ -120,6 +120,7 @@ def __init__(
         feature_name: Union[List[str], str] = "auto",
         feval: Optional[Callable] = None,
         fobj: Optional[Callable] = None,
+        init_model: Optional[Union[lgb.Booster, lgb.LGBMModel, str]] = None,
         n_estimators: int = 100,
         param_distributions: Optional[
             Dict[str, distributions.BaseDistribution]
@@ -135,6 +136,7 @@ def __init__(
         self.feature_name = feature_name
         self.feval = feval
         self.fobj = fobj
+        self.init_model = init_model
         self.is_higher_better = is_higher_better
         self.n_estimators = n_estimators
         self.n_samples = n_samples
@@ -155,6 +157,7 @@ def __call__(self, trial: trial_module.Trial) -> float:
             feval=self.feval,
             fobj=self.fobj,
             folds=self.cv,
+            init_model=self.init_model,
             num_boost_round=self.n_estimators,
         )  # Dict[str, List[float]]
         best_iteration = callbacks[0]._best_iteration  # type: ignore
@@ -386,6 +389,7 @@ def _make_booster(
         feature_name: Union[List[str], str] = "auto",
         categorical_feature: Union[List[int], List[str], str] = "auto",
         callbacks: Optional[List[Callable]] = None,
+        init_model: Optional[Union[lgb.Booster, lgb.LGBMModel, str]] = None,
     ) -> Union[_VotingBooster, lgb.Booster]:
         if self.refit:
             booster = lgb.train(
@@ -396,6 +400,7 @@ def _make_booster(
                 feature_name=feature_name,
                 categorical_feature=categorical_feature,
                 callbacks=callbacks,
+                init_model=init_model,
             )
 
             booster.free_dataset()
@@ -424,6 +429,7 @@ def fit(
         feature_name: Union[List[str], str] = "auto",
         categorical_feature: Union[List[int], List[str], str] = "auto",
         callbacks: Optional[List[Callable]] = None,
+        init_model: Optional[Union[lgb.Booster, lgb.LGBMModel, str]] = None,
         groups: Optional[OneDimArrayLikeType] = None,
         **fit_params: Any
     ) -> "LGBMModel":
@@ -467,6 +473,10 @@ def fit(
         callbacks
             List of callback functions that are applied at each iteration.
 
+        init_model
+            Filename of LightGBM model, Booster instance or LGBMModel instance
+            used for continue training.
+
         groups
             Group labels for the samples used while splitting the dataset into
             train/test set. If `group` is not None, this parameter is ignored.
@@ -542,6 +552,9 @@ def fit(
             eval_name = params["metric"]
             is_higher_better = _is_higher_better(params["metric"])
 
+        if isinstance(init_model, lgb.LGBMModel):
+            init_model = init_model.booster_
+
         if self.study is None:
             sampler = samplers.TPESampler(seed=seed)
 
@@ -578,6 +591,7 @@ def fit(
             feature_name=feature_name,
             feval=feval,
             fobj=fobj,
+            init_model=init_model,
             n_estimators=self.n_estimators,
             param_distributions=self.param_distributions,
         )
@@ -841,6 +855,7 @@ def fit(
         feature_name: Union[List[str], str] = "auto",
         categorical_feature: Union[List[int], List[str], str] = "auto",
         callbacks: Optional[List[Callable]] = None,
+        init_model: Optional[Union[lgb.Booster, lgb.LGBMModel, str]] = None,
         groups: Optional[OneDimArrayLikeType] = None,
         **fit_params: Any
     ) -> "LGBMClassifier":
@@ -884,6 +899,10 @@ def fit(
         callbacks
             List of callback functions that are applied at each iteration.
 
+        init_model
+            Filename of LightGBM model, Booster instance or LGBMModel instance
+            used for continue training.
+
         groups
             Group labels for the samples used while splitting the dataset into
             train/test set. If `group` is not None, this parameter is ignored.
@@ -913,6 +932,7 @@ def fit(
             feature_name=feature_name,
             categorical_feature=categorical_feature,
             callbacks=callbacks,
+            init_model=init_model,
             groups=groups,
             **fit_params
         )