Update app.py

app.py

@@ -51,7 +51,7 @@ def synthesize_speech(text, sentence_silence, length_scale, normalize_text=True)
     audio_data = np.frombuffer(buffer.read(), dtype=np.int16)
     inference_time = time.time() - start
     metric_text = f"Time to generate audio: {round(inference_time*1000)} milliseconds\n"
-    return (voice.config.sample_rate, audio_data), "{}s".format(metric_text)
+    return (voice.config.sample_rate, audio_data), metric_text
 
 
 with gr.Blocks(analytics_enabled=False) as demo: