diff --git a/examples/waveflow/synthesis.py b/examples/waveflow/synthesis.py
index 15c4d3b843165540c6f80f986b23e73ffeb4a59e..b9569bfcfae3f186852e5c36ecec05d3f668b6c9 100644
--- a/examples/waveflow/synthesis.py
+++ b/examples/waveflow/synthesis.py
@@ -93,16 +93,7 @@ def synthesize(config):
 
         # Build model.
         model = WaveFlow(config, checkpoint_dir)
-        model.build(training=False)
-        # Obtain the current iteration.
-        if config.checkpoint is None:
-            if config.iteration is None:
-                iteration = io.load_latest_checkpoint(checkpoint_dir)
-            else:
-                iteration = config.iteration
-        else:
-            iteration = int(config.checkpoint.split('/')[-1].split('-')[-1])
-
+        iteration = model.build(training=False)
         # Run model inference.
         model.infer(iteration)
 
diff --git a/examples/waveflow/waveflow.py b/examples/waveflow/waveflow.py
index 700116b4f2bb33d764acb759aee68c8aa9827162..23c558ec74a4b9ff81541beed4eb4075ae791440 100644
--- a/examples/waveflow/waveflow.py
+++ b/examples/waveflow/waveflow.py
@@ -81,12 +81,6 @@ class WaveFlow():
 
         waveflow = WaveFlowModule(config)
 
-        # Dry run once to create and initalize all necessary parameters.
-        audio = dg.to_variable(np.random.randn(1, 16000).astype(self.dtype))
-        mel = dg.to_variable(
-            np.random.randn(1, config.mel_bands, 63).astype(self.dtype))
-        waveflow(audio, mel)
-
         if training:
             optimizer = fluid.optimizer.AdamOptimizer(
                 learning_rate=config.learning_rate,