update

app.py

@@ -35,7 +35,7 @@ def tts_inference(text, prompt_wav='examples/prompt_1.wav', lang='zh'):
         prompt_wav=prompt_wav,
         text=text,
         lang=lang,
-    ).detach().cpu().numpy()
+    )[0].detach().cpu().numpy()
     print(f'Generate waveform with the shape of {syn_audio.shape}')
     syn_audio = (syn_audio * 32768).astype(np.int16)
     return 48000, syn_audio