fix

app.py

@@ -72,8 +72,11 @@ def generate_answer(image, question):
     return answer_ru
 
 
-def transcribe(image, new_chunk):
-    sr, y = new_chunk
+def transcribe(image, audio):
+    if not image or not audio:
+        return
+
+    sr, y = audio
 
     # Convert to mono if stereo
     if y.ndim > 1:
@@ -81,7 +84,6 @@ def transcribe(image, new_chunk):
 
     y = y.astype(np.float32)
     y /= np.max(np.abs(y))
-
     return generate_answer(image, transcriber({"sampling_rate": sr, "raw": y})["text"])