diff --git a/data_utils/data.py b/data_utils/data.py
index 0fb2a88b..a526da91 100644
--- a/data_utils/data.py
+++ b/data_utils/data.py
@@ -330,11 +330,10 @@ class DataGenerator(object):
                 axis=0)
             masks.append(mask)
         padded_audios = np.array(padded_audios).astype('float32')
+        texts = np.expand_dims(np.array(texts).astype('int32'), axis=-1)
         if self._is_training:
             texts = fluid.create_lod_tensor(
-                np.array(texts).astype('int32'),
-                recursive_seq_lens=[text_lens],
-                place=self._place)
+                texts, recursive_seq_lens=[text_lens], place=self._place)
         audio_lens = np.array(audio_lens).astype('int64').reshape([-1, 1])
         masks = np.array(masks).astype('float32')
         return padded_audios, texts, audio_lens, masks