Fix: Add default parameters to generate_response function

app.py

@@ -102,7 +102,7 @@ def get_duration(prompt, max_new_tokens, temperature, top_p):
     return base_duration + token_duration
 
 @spaces.GPU(duration=get_duration)
-def generate_response(prompt, max_new_tokens, temperature, top_p):
+def generate_response(prompt, max_new_tokens=MAX_NEW_TOKENS, temperature=DEFAULT_TEMPERATURE, top_p=DEFAULT_TOP_P):
     """
     Generate response using the router model.
     In ZeroGPU mode: model is loaded on CPU, moved to GPU here, then back to CPU after.