Update app.py

app.py

@@ -32,11 +32,8 @@ def tts_mode(txt, voice):
     raw_path.seek(0)
     model = Svc(fr"Herta-Svc/G_10000.pth", f"Herta-Svc/config.json", device = 'cpu')
     out_audio, out_sr = model.infer('speaker0', 0, raw_path, auto_predict_f0 = True,)
-    soundfile.write('xxx.wav', out_audio.cpu().numpy(), 44100)
-    audio_file = open('xxx.wav', 'rb')
-    audio_bytes = audio_file.read()
-
-    return audio_bytes
+   
+    return (44100, out_audio.cpu().numpy())
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()