add max duration of 2 minutes and verbose llamacpp

app.py

@@ -71,14 +71,14 @@ CREATE TABLE {} (
     return sql_ddl
 
 
-@spaces.GPU
+@spaces.GPU(duration=120)
 def generate_query(dataset_id: str, query: str) -> dict:
     llama = llama_cpp.Llama(
         model_path="Hermes-2-Pro-Llama-3-8B-Q8_0.gguf",
         n_gpu_layers=50,
         chat_format="chatml",
         n_ctx=2048,
-        verbose=False,
+        verbose=True,
         temperature=0.1,
     )