diff --git a/modelscope/msdatasets/ms_dataset.py b/modelscope/msdatasets/ms_dataset.py
index ad900bab..e90f397b 100644
--- a/modelscope/msdatasets/ms_dataset.py
+++ b/modelscope/msdatasets/ms_dataset.py
@@ -563,6 +563,18 @@ class MsDataset:
         self._hf_ds.reset_format()
         return self._hf_ds
 
+    def remap_columns(self, column_mapping: Dict[str, str]) -> Dataset:
+        """
+        Rename columns and return the underlying hf dataset directly
+        TODO: support native MsDataset column rename.
+        Args:
+            column_mapping: the mapping of the original and new column names
+        Returns:
+            underlying hf dataset
+        """
+        self._hf_ds.reset_format()
+        return self._hf_ds.rename_columns(column_mapping)
+
     @staticmethod
     def upload(object_name: str,
                local_file_path: str,
diff --git a/tests/trainers/test_finetune_mplug.py b/tests/trainers/test_finetune_mplug.py
index 72196fba..4972a731 100644
--- a/tests/trainers/test_finetune_mplug.py
+++ b/tests/trainers/test_finetune_mplug.py
@@ -24,17 +24,16 @@ class TestFinetuneMPlug(unittest.TestCase):
         datadict = MsDataset.load(
             'coco_captions_small_slice',
             download_mode=DownloadMode.FORCE_REDOWNLOAD)
-        self.train_dataset = MsDataset(datadict['train'].to_hf_dataset().map(
-            lambda _: {
-                'question': 'what the picture describes?'
-            }).rename_column('image:FILE',
-                             'image').rename_column('answer:Value', 'answer'))
-        self.test_dataset = MsDataset(datadict['test'].to_hf_dataset().map(
-            lambda _: {
-                'question': 'what the picture describes?'
-            }).rename_column('image:FILE',
-                             'image').rename_column('answer:Value', 'answer'))
-
+        self.train_dataset = MsDataset(
+            datadict['train'].remap_columns({
+                'image:FILE': 'image',
+                'answer:Value': 'answer'
+            }).map(lambda _: {'question': 'what the picture describes?'}))
+        self.test_dataset = MsDataset(
+            datadict['test'].remap_columns({
+                'image:FILE': 'image',
+                'answer:Value': 'answer'
+            }).map(lambda _: {'question': 'what the picture describes?'}))
         self.max_epochs = 2
 
     def tearDown(self):
diff --git a/tests/trainers/test_finetune_text_generation.py b/tests/trainers/test_finetune_text_generation.py
index 63d4577b..59bef51c 100644
--- a/tests/trainers/test_finetune_text_generation.py
+++ b/tests/trainers/test_finetune_text_generation.py
@@ -130,10 +130,16 @@ class TestFinetuneTextGeneration(unittest.TestCase):
     def test_finetune_cnndm(self):
         from modelscope.msdatasets import MsDataset
         dataset_dict = MsDataset.load('DuReader_robust-QG')
-        train_dataset = dataset_dict['train'].to_hf_dataset() \
-            .rename_columns({'text1': 'src_txt', 'text2': 'tgt_txt'})
-        eval_dataset = dataset_dict['validation'].to_hf_dataset() \
-            .rename_columns({'text1': 'src_txt', 'text2': 'tgt_txt'})
+        train_dataset = dataset_dict['train'].remap_columns({
+            'text1': 'src_txt',
+            'text2': 'tgt_txt'
+        })
+        eval_dataset = dataset_dict['validation'].remap_columns({
+            'text1':
+            'src_txt',
+            'text2':
+            'tgt_txt'
+        })
         num_warmup_steps = 200
         os.environ['LOCAL_RANK'] = '0'