Fix fit on LGBMModel

optgbm/sklearn.py

@@ -522,21 +522,13 @@ def fit(
         ):
             params.pop(attr, None)
 
+        params["objective"] = self._get_objective()
         params["random_state"] = seed
         params["verbose"] = -1
 
         if self._n_classes is not None and self._n_classes > 2:
             params["num_classes"] = self._n_classes
 
-        if callable(self.objective):
-            fobj = _ObjectiveFunctionWrapper(self.objective)
-        else:
-            fobj = None
-
-        self._objective = self._get_objective()
-
-        params["objective"] = self._objective
-
         if callable(eval_metric):
             params["metric"] = "None"
             feval = _EvalFunctionWrapper(eval_metric)
@@ -552,6 +544,11 @@ def fit(
             eval_name = params["metric"]
             is_higher_better = _is_higher_better(params["metric"])
 
+        if callable(self.objective):
+            fobj = _ObjectiveFunctionWrapper(self.objective)
+        else:
+            fobj = None
+
         if isinstance(init_model, lgb.LGBMModel):
             init_model = init_model.booster_
 
@@ -612,6 +609,7 @@ def fit(
             "best_iteration"
         ]
         self._best_score = self.study_.best_value
+        self._objective = params["objective"]
         self.best_params_ = {**params, **self.study_.best_params}
         self.n_splits_ = cv.get_n_splits(X, y, groups=groups)