Update app.py

app.py

@@ -38,29 +38,25 @@ def speech_to_text(audio_data, tgt_lang):
     audio_input, _ = torchaudio.load(file_path)
     s2t_model = torch.jit.load("unity_on_device.ptl")
     with torch.no_grad():
-        text = s2t_model(audio_input, tgt_lang=languages[tgt_lang])
+        model_output = s2t_model(audio_input, tgt_lang=languages[tgt_lang])
+    transcribed_text = model_output[0] if model_output else ""
+    print("Speech to Text Model Output:", transcribed_text)
 
-    # Print the model's output for debugging
-    print("Speech to Text Model Output:", text)
-
-    return text
+    return transcribed_text
 
 def speech_to_speech_translation(audio_data, tgt_lang):
     file_path = save_audio(audio_data)
     audio_input, _ = torchaudio.load(file_path)
     s2st_model = torch.jit.load("unity_on_device.ptl")
-
     with torch.no_grad():
-        text, units, waveform = s2st_model(audio_input, tgt_lang=languages[tgt_lang])
-
-    # Print the model's output for debugging (optional)
-    print("Translated Text:", text)
+        translated_text, units, waveform = s2st_model(audio_input, tgt_lang=languages[tgt_lang])
+    output_file = "/tmp/result.wav"
+    torchaudio.save(output_file, waveform.unsqueeze(0), sample_rate=16000)
+    print("Translated Text:", translated_text)
     print("Units:", units)
     print("Waveform Shape:", waveform.shape)
 
-    output_file = "/tmp/result.wav"
-    torchaudio.save(output_file, waveform.unsqueeze(0), sample_rate=16000)
-    return text, output_file
+    return translated_text, output_file
 
 
 def create_interface():