Update app.py

app.py

@@ -170,7 +170,7 @@ def load_models_and_documents():
             quantization="bitsandbytes",
             load_format="bitsandbytes",
             gpu_memory_utilization=0.9,
-            max_model_len=12288,
+            max_model_len=12000,
             dtype=torch.float16
         )
         model = models.VLLM(llm)