Pass correct split_type to GPU predictor

src/predictor/gpu_predictor.cu

@@ -272,7 +272,11 @@ PredictKernel(Data data, common::Span<const RegTree::Node> d_nodes,
           d_categories.subspan(d_cat_tree_segments[tree_idx - tree_begin],
                                d_cat_tree_segments[tree_idx - tree_begin + 1] -
                                d_cat_tree_segments[tree_idx - tree_begin]);
-      float leaf = GetLeafWeight(global_idx, d_tree, d_tree_split_types,
+      auto tree_split_types =
+          d_tree_split_types.subspan(d_tree_segments[tree_idx - tree_begin],
+                                     d_tree_segments[tree_idx - tree_begin + 1] -
+                                     d_tree_segments[tree_idx - tree_begin]);
+      float leaf = GetLeafWeight(global_idx, d_tree, tree_split_types,
                                  tree_cat_ptrs,
                                  tree_categories,
                                  &loader);

tests/python-gpu/test_gpu_prediction.py

@@ -3,8 +3,17 @@
 
 import numpy as np
 import xgboost as xgb
+from xgboost.compat import PANDAS_INSTALLED
+
 from hypothesis import given, strategies, assume, settings, note
 
+if PANDAS_INSTALLED:
+    from hypothesis.extra.pandas import column, data_frames, range_indexes
+else:
+    def noop(*args, **kwargs):
+        pass
+    column, data_frames, range_indexes = noop, noop, noop
+
 sys.path.append("tests/python")
 import testing as tm
 from test_predict import run_threaded_predict  # noqa
@@ -259,3 +268,29 @@ def test_predict_leaf_dart(self, param, dataset):
         param['booster'] = 'dart'
         param['tree_method'] = 'gpu_hist'
         self.run_predict_leaf_booster(param, 10, dataset)
+
+    @pytest.mark.skipif(**tm.no_sklearn())
+    @pytest.mark.skipif(**tm.no_pandas())
+    @given(df=data_frames([column('x0', elements=strategies.integers(min_value=0, max_value=3)),
+                           column('x1', elements=strategies.integers(min_value=0, max_value=5))],
+                          index=range_indexes(min_size=20, max_size=50)))
+    @settings(deadline=None)
+    def test_predict_categorical_split(self, df):
+        from sklearn.metrics import mean_squared_error
+
+        df = df.astype('category')
+        x0, x1 = df['x0'].to_numpy(), df['x1'].to_numpy()
+        y = (x0 * 10 - 20) + (x1 - 2)
+        dtrain = xgb.DMatrix(df, label=y, enable_categorical=True)
+
+        params = {'tree_method': 'gpu_hist', 'predictor': 'gpu_predictor',
+                  'enable_experimental_json_serialization': True,
+                  'max_depth': 3, 'learning_rate': 1.0, 'base_score': 0.0, 'eval_metric': 'rmse'}
+
+        eval_history = {}
+        bst = xgb.train(params, dtrain, num_boost_round=5, evals=[(dtrain, 'train')],
+                        verbose_eval=False, evals_result=eval_history)
+
+        pred = bst.predict(dtrain)
+        rmse = mean_squared_error(y_true=y, y_pred=pred, squared=False)
+        np.testing.assert_almost_equal(rmse, eval_history['train']['rmse'][-1], decimal=5)