Update app.py

Adjusting down max tokens to work with the overarching token restrictions

app.py

@@ -155,7 +155,7 @@ final_answer = FinalAnswerTool()
 
 model = HfApiModel(
 #max_tokens=2096,
-max_tokens=20960,
+max_tokens=8000,
 #max_tokens=256,
 temperature=0.5,
 model_id='Qwen/Qwen2.5-Coder-32B-Instruct',# it is possible that this model may be overloaded