Update app.py

app.py

@@ -149,7 +149,7 @@ def get_response(user_input):
     llm = LlamaCpp(
     model_path="tinyllama-1.1b-chat-v1.0.Q4_0.gguf",
     temperature=0.75,
-    max_tokens=2000,
+    max_tokens=500,
     top_p=1,
     # callback_manager=callback_manager,
     # verbose=True,  # Verbose is required to pass to the callback manager