diff --git a/paddlespeech/server/engine/tts/paddleinference/tts_engine.py b/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
index 2878c852..20b98fae 100644
--- a/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
+++ b/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
@@ -247,11 +247,13 @@ class TTSServerExecutor(TTSExecutor):
             else:
                 # multi speaker  do not have static model
                 if am_dataset in {"aishell3", "vctk"}:
-                    pass
+                    am_result = run_model(
+                        self.am_predictor,
+                        [part_phone_ids.numpy(), np.array([spk_id])])
                 else:
                     am_result = run_model(self.am_predictor,
                                           [part_phone_ids.numpy()])
-                    mel = am_result[0]
+                mel = am_result[0]
             self.am_time += (time.time() - am_st)
 
             # voc