Update app.py

app.py

@@ -25,7 +25,7 @@ def tts_get_voices_list():
     
     return voices
 
-def tts_mode(txt, voice):
+def infer(txt, input_audio, voice, audio_mode):
     tts = asyncio.run(edge_tts.Communicate(txt, voice).save('audio.mp3'))
     audio, sr = librosa.load('audio.mp3', sr=16000, mono=True)
     raw_path = io.BytesIO()
@@ -33,23 +33,21 @@ 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,)
-   
-    return (44100, out_audio.cpu().numpy())
-
-def audio_infer_mode(input_audio):
-    sampling_rate, audio = input_audio
-    audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
     
-    if len(audio.shape) > 1:
-        audio = librosa.to_mono(audio.transpose(1, 0))
-    if sampling_rate != 16000:
-        audio = librosa.resample(audio, org_sr=sampling_rate, target_sr=16000)
+    if audio_mode:
+        sampling_rate, audio = input_audio
+        audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
+    
+        if len(audio.shape) > 1:
+            audio = librosa.to_mono(audio.transpose(1, 0))
+        if sampling_rate != 16000:
+            audio = librosa.resample(audio, org_sr=sampling_rate, target_sr=16000)
 
-    raw_path = io.BytesIO()
-    soundfile.write(raw_path, audio, 16000, format="wav")
-    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,)
+        raw_path = io.BytesIO()
+        soundfile.write(raw_path, audio, 16000, format="wav")
+        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,)
 
     return (44100, out_audio.cpu().numpy())
 
@@ -84,6 +82,5 @@ if __name__ == '__main__':
                 tts_voice = gr.Dropdown.update(choices= tts_get_voices_list(), visible = True, show_label = True)
                 audio_input = gr.Audio.update(label = 'Input Audio', visible = False, show_label = False)
 
-        btn_submit.click([tts_mode, audio_infer_mode], [tts_text, tts_voice, audio_input], [audio_output])
-
+        btn_submit.click(infer, [tts_text, audio_input, tts_voice], [audio_output])
     app.queue(concurrency_count=1, api_open=args.api).launch(share=args.share)