Update app.py

app.py

@@ -14,23 +14,25 @@ tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device)
 @spaces.GPU(enable_queue=True)
 def clone(text, audio):
     # Generowanie mowy
-    wav, alignment, text_info, _ = tts.tts(text=text, speaker_wav=audio, language="pl", return_type="dict")
+    wav = tts.tts(text=text, speaker_wav=audio, language="pl")
     
     # Zapisywanie pliku audio
     tts.save_wav(wav, file_path="./output.wav")
     
+    # Uzyskanie informacji o fonemach
+    phonemes_info = tts.synthesizer.get_phonemes(text, language="pl")
+    
     # Przygotowanie informacji o fonemach
-    phonemes_info = []
-    for phoneme, start, end in zip(text_info["phonemes"], alignment["align_durations"], alignment["align_durations_cumsum"]):
-        phonemes_info.append({
+    phonemes_data = []
+    for phoneme, duration in phonemes_info:
+        phonemes_data.append({
             "phoneme": phoneme,
-            "start": float(start),
-            "end": float(end)
+            "duration": float(duration)
         })
     
     # Zapisywanie informacji o fonemach do pliku JSON
     with open("./phonemes_info.json", "w", encoding="utf-8") as f:
-        json.dump(phonemes_info, f, ensure_ascii=False, indent=2)
+        json.dump(phonemes_data, f, ensure_ascii=False, indent=2)
     
     return "./output.wav", "./phonemes_info.json"
 
@@ -49,4 +51,4 @@ iface = gr.Interface(
     theme=gr.themes.Base(primary_hue="teal", secondary_hue="teal", neutral_hue="slate")
 )
 
-iface.launch()
+iface.launch(share=True)