Update xtts.py

xtts.py

@@ -150,4 +150,4 @@ def trim_sample_audio(speaker_wav):
         traceback.print_exc()
         return speaker_wav
 
-gr.Inference(predict, inputs=["text", "text"], outputs="text").launch()
+gr.Interface(predict, inputs=["text", "text"], outputs=gr.Audio).launch()