Update app.py

app.py

@@ -9,7 +9,7 @@ client = InferenceClient("microsoft/Phi-3-mini-4k-instruct")
 def respond(
         message,
         history: list[tuple[str, str]],
-        max_tokens=1024,
+        max_tokens=512,
         temperature=0.7,
         top_p=0.95,
 ):