diff --git a/paddlespeech/t2s/exps/ort_predict.py b/paddlespeech/t2s/exps/ort_predict.py
index 271e6d0ddec079648fad4c94d715e6c3fbb3b371..a55713366b2662e1c9617a8b136d1fc5c09724d3 100644
--- a/paddlespeech/t2s/exps/ort_predict.py
+++ b/paddlespeech/t2s/exps/ort_predict.py
@@ -36,7 +36,7 @@ def get_sess(args, filed='am'):
     sess_options.execution_mode = ort.ExecutionMode.ORT_SEQUENTIAL
 
     if args.device == "gpu":
-        # fastspeech2 can't use trt now!
+        # fastspeech2/mb_melgan can't use trt now!
         if args.use_trt:
             providers = ['TensorrtExecutionProvider']
         else:
diff --git a/paddlespeech/t2s/exps/ort_predict_e2e.py b/paddlespeech/t2s/exps/ort_predict_e2e.py
index 5b5d976860ece10c3f23474bea162904bec1b16d..a52085679ceb808760b72bd063c2e6772a59db9a 100644
--- a/paddlespeech/t2s/exps/ort_predict_e2e.py
+++ b/paddlespeech/t2s/exps/ort_predict_e2e.py
@@ -36,7 +36,7 @@ def get_sess(args, filed='am'):
     sess_options.execution_mode = ort.ExecutionMode.ORT_SEQUENTIAL
 
     if args.device == "gpu":
-        # fastspeech2 can't use trt now!
+        # fastspeech2/mb_melgan can't use trt now!
         if args.use_trt:
             providers = ['TensorrtExecutionProvider']
         else: