Update app.py

app.py

@@ -39,16 +39,9 @@ def parse_multilingual_text(input_text):
 
 @spaces.GPU
 def generate_segment_audio(text, lang, speaker_audio, pipe):
-    speaker_audio_data = None
     if not isinstance(text, str):
         text = text.decode("utf-8") if isinstance(text, bytes) else str(text)
-    if speaker_audio is not None:
-        if isinstance(speaker_audio, str):
-            with open(speaker_audio, 'rb') as file:
-                speaker_audio_data = file.read()
-        else:
-            speaker_audio_data = speaker_audio
-        speaker_audio_data = torch.compile.tensor(speaker_audio_data)
+    speaker_audio_data = speaker_audio
     audio_data = pipe.generate(text, speaker_audio_data, lang)
     resample_audio = resampler(newsr=24000)
     audio_data_resampled = next(resample_audio([{'sample_rate': 24000, 'samples': audio_data.cpu()}]))['samples_24k']