diff --git a/paddlespeech/t2s/models/fastspeech2/fastspeech2.py b/paddlespeech/t2s/models/fastspeech2/fastspeech2.py
index 0eb44beb6d231c1a1e55e9472ff3445ca26676fb..c790c8cb2a6452c1c151de83a2490cd7fa68ba1b 100644
--- a/paddlespeech/t2s/models/fastspeech2/fastspeech2.py
+++ b/paddlespeech/t2s/models/fastspeech2/fastspeech2.py
@@ -705,9 +705,9 @@ class FastSpeech2(nn.Layer):
     def encoder_infer(
             self,
             text: paddle.Tensor,
+            spk_id=None,
             alpha: float=1.0,
             spk_emb=None,
-            spk_id=None,
             tone_id=None,
     ) -> Tuple[paddle.Tensor, paddle.Tensor, paddle.Tensor]:
         # input of embedding must be int64