Update app.py

app.py

@@ -3,6 +3,7 @@ from llama_cpp import Llama
 import os
 from groq import Groq
 import numpy as np
+from scipy.io import wavfile
 
 #tts
 from balacoon_tts import TTS
@@ -46,9 +47,12 @@ llm = Llama.from_pretrained(
 def text_to_speech(text):
     with tempfile.NamedTemporaryFile(delete=False, suffix='.wav') as temp_file:
         with locker:
-            audio_data = tts.synthesize(text, "92")
-        temp_file.write(np.ascontiguousarray(audio_data))
-    return temp_file.name
+            audio_data, sample_rate = tts.synthesize(text, "92")
+            audio_data_int16 = (audio_data * 32767).astype(np.int16)
+        output_file = temp_file.name
+        wavfile.write(output_file, sample_rate, audio_data_int16)
+    return output_file
+    
 
 def combine_audio_files(audio_files):
     combined = AudioSegment.empty()