fix xgb losses and metrics

lightautoml/ml_algo/boost_xgb.py

@@ -87,10 +87,10 @@ def _infer_params(
         fobj = loss.fobj
 
         # # get metric params
-        # params["metric"] = loss.metric_name
+        params["eval_metric"] = loss.metric_name
         feval = loss.feval
 
-        # params["num_class"] = self.n_classes
+        params["num_class"] = self.n_classes
         # add loss and tasks params if defined
         params = {**params, **loss.fobj_params, **loss.metric_params}
 
@@ -188,24 +188,24 @@ def _get_default_search_spaces(self, suggested_params: Dict, estimated_n_trials:
         """
         optimization_search_space = {}
 
-        optimization_search_space["feature_fraction"] = Uniform(
+        optimization_search_space["colsample_bytree"] = Uniform(
             low=0.5,
             high=1.0,
         )
 
-        optimization_search_space["num_leaves"] = Uniform(
+        optimization_search_space["max_leaves"] = Uniform(
             low=16,
             high=255,
             q=1,
         )
 
         if estimated_n_trials > 30:
-            optimization_search_space["bagging_fraction"] = Uniform(
+            optimization_search_space["subsample"] = Uniform(
                 low=0.5,
                 high=1.0,
             )
 
-            optimization_search_space["min_sum_hessian_in_leaf"] = Uniform(
+            optimization_search_space["min_child_weight"] = Uniform(
                 low=1e-3,
                 high=10.0,
                 log=True,

lightautoml/tasks/losses/xgb.py

@@ -30,14 +30,14 @@
 }
 
 _xgb_reg_metrics_dict = {
-    "mse": "mse",  # TODO
+    "mse": "rmse",
     "mae": "mae",
-    # "r2": "mse", # TODO
+    "r2": "rmse",
     "rmsle": "rmsle",
     "mape": "mape",
-    "quantile": "reg:quantileerror",
-    "huber": "reg:pseudohubererror",
-    # "fair": "fair",# TODO
+    # "quantile": "", # TODO
+    "huber": "mphe",
+    # "fair": "",# TODO
 }
 
 _xgb_multiclass_metrics_dict = {

pyproject.toml

@@ -47,6 +47,7 @@ scikit-learn = [
 ]
 lightgbm = ">=2.3, <=3.2.1"
 catboost = ">=0.26.1"
+xgboost = "^2.0.0"
 optuna = "*"
 torch = [
     {platform = "win32", python = "3.6.1", version = "1.7.0"},