Update app.py

app.py

@@ -69,9 +69,8 @@ def infer(sample_audio_path, target_text, progress=gr.Progress()):
             # If already mono, just use the original waveform
             waveform_mono = waveform
         prompt_wav = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)(waveform_mono)
-        prompt_text = whisper_turbo_pipe(prompt_wav[0].cpu().numpy(), language="en")['text'].strip()  # ✅ Force English transcription
+        prompt_text = whisper_turbo_pipe(prompt_wav[0].numpy())['text'].strip()
         progress(0.5, 'Transcribed! Generating speech...')
-
         if len(target_text) == 0:
             return None
         elif len(target_text) > 300: