Update app.py

app.py

@@ -129,8 +129,12 @@ def text_to_speech(text, audio_file=None):
     # Convert the generated speech to numpy array format
     speech_np = speech.cpu().numpy()
     
-    # Return the numpy array and the sample rate
-    return (speech_np, 16000)
+    # Save the speech to a temporary file in WAV format
+    output_file = "output.wav"
+    sf.write(output_file, speech_np, 16000)
+
+    # Return the path to the audio file
+    return output_file
 
 iface = gr.Interface(
     fn=text_to_speech,