CatBoostError: catboost/libs/data/model_dataset_compatibility.cpp:53: Feature feature_49 在模型中是分类的，但在数据集中标记不同

如何解决CatBoostError: catboost/libs/data/model_dataset_compatibility.cpp:53: Feature feature_49 在模型中是分类的，但在数据集中标记不同

我第一次使用 catboost，但是当我尝试通过 predict_proba 预测类别概率时出现错误，我仅在设置 iterations > 50 时才会收到此错误,y_train 的值是 float 并且 X_train 包含 50 个特征并且都有 int 值，cat_features 是一个列表 range 0 to 49，

我的代码：

from catboost import CatBoostClassifier
clf = CatBoostClassifier(
    iterations=200,random_seed=42,learning_rate=0.1,custom_loss=['AUC','Accuracy']
)

clf.fit(
    X_train,y_train,cat_features=cat_features,eval_set=(X_val,y_val),verbose=1,plot=True,)

错误信息：

---

CatBoostError                             Traceback (most recent call last)
<ipython-input-64-be92151f9ec3> in <module>
----> 1 predict = clf.predict_proba(test_)

/opt/conda/lib/python3.7/site-packages/catboost/core.py in predict_proba(self,data,ntree_start,ntree_end,thread_count,verbose)
   4631                 with probability for every class for each object.
   4632         """
-> 4633         return self._predict(data,'Probability',verbose,'predict_proba')
   4634 
   4635 

/opt/conda/lib/python3.7/site-packages/catboost/core.py in _predict(self,prediction_type,parent_method_name)
   2091         self._validate_prediction_type(prediction_type)
   2092 
-> 2093         predictions = self._base_predict(data,verbose)
   2094         return predictions[0] if data_is_single_object else predictions
   2095 

/opt/conda/lib/python3.7/site-packages/catboost/core.py in _base_predict(self,pool,verbose)
   1416 
   1417     def _base_predict(self,verbose):
-> 1418         return self._object._base_predict(pool,verbose)
   1419 
   1420     def _base_virtual_ensembles_predict(self,virtual_ensembles_count,verbose):

_catboost.pyx in _catboost._CatBoost._base_predict()

_catboost.pyx in _catboost._CatBoost._base_predict()

CatBoostError: catboost/libs/data/model_dataset_compatibility.cpp:53: Feature feature_49 is Categorical in model but marked different in the dataset