[BUG] test_gpu_treeshap.py::test_xgb_regressor[reg:pseudohubererror] fails with XGBoost 1.6.0

[hcho3]

A pytest fails with the latest XGBoost (1.6.0).

Reproduction:

python -m pytest -v -x -rxXs python/cuml/tests/explainer/test_gpu_treeshap.py

Error message:

python/cuml/tests/explainer/test_gpu_treeshap.py::test_xgb_regressor[reg:pseudohubererror] FAILED                                                                      [ 10%] 
                                                                                                                                                                              
================================================================================== FAILURES ==================================================================================
__________________________________________________________________ test_xgb_regressor[reg:pseudohubererror] __________________________________________________________________
                                                                                                                                                                              
objective = 'reg:pseudohubererror'                                                                                                                                            
                                                                                                                                                                              
    @pytest.mark.parametrize('objective', ['reg:linear', 'reg:squarederror',                                                                                                  
                                           'reg:squaredlogerror',                                                                                                             
                                           'reg:pseudohubererror'])                                                                                                           
    @pytest.mark.skipif(not has_xgboost(), reason="need to install xgboost")                                                                                                  
    @pytest.mark.skipif(not has_shap(), reason="need to install shap")                                                                                                        
    @pytest.mark.skipif(not has_sklearn(), reason="need to install scikit-learn")                                                                                             
    def test_xgb_regressor(objective):                                                                                                                                        
        n_samples = 100                                                                                                                                                       
        X, y = make_regression(n_samples=n_samples, n_features=8, n_informative=8,                                                                                            
                               n_targets=1, random_state=2021)                                                                                                                
        # Ensure that the label exceeds -1                                                                                                                                    
        y += (-0.5) - np.min(y)                                                                                                                                               
        assert np.all(y > -1)                                                                                                                                                 
        X, y = X.astype(np.float32), y.astype(np.float32)                                                                                                                     
        dtrain = xgb.DMatrix(X, label=y)
        params = {'objective': objective, 'base_score': 0.5, 'seed': 0,
                  'max_depth': 6, 'tree_method': 'gpu_hist',
                  'predictor': 'gpu_predictor'}
        num_round = 10
        xgb_model = xgb.train(params, dtrain, num_boost_round=num_round,
                              evals=[(dtrain, 'train')])
>       tl_model = treelite.Model.from_xgboost(xgb_model)

python/cuml/tests/explainer/test_gpu_treeshap.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
../../miniconda3/envs/cuml_dev/lib/python3.9/site-packages/treelite/frontend.py:388: in from_xgboost
    return cls.from_xgboost_json(model_json_str)
../../miniconda3/envs/cuml_dev/lib/python3.9/site-packages/treelite/frontend.py:436: in from_xgboost_json
    _check_call(_LIB.TreeliteLoadXGBoostJSONString(
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

ret = -1

    def _check_call(ret):
		"""Check the return value of C API call
     
        This function will raise exception when error occurs.
        Wrap every API call with this function
     
        Parameters
        ----------
        ret : int
            return value from API calls
        """
        if ret != 0:
>           raise TreeliteError(_LIB.TreeliteGetLastError().decode('utf-8'))
E           treelite.util.TreeliteError: [18:17:57] /home/conda/feedstock_root/build_artifacts/treelite_1645138081934/work/src/frontend/xgboost_json.cc:539: Provided JSON could not be parsed as XGBoost model. Parsing error at offset 4458: Terminate parsing due to Handler error.
E           ame":"reg:pseudohubererror","pseduo_huber_param":{"huber_slope":"1"}}},"version":[1,6,0]}
E           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

../../miniconda3/envs/cuml_dev/lib/python3.9/site-packages/treelite/core.py:50: TreeliteError

[github-actions]

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.