Update app.py

app.py

@@ -185,7 +185,7 @@ def load_models_and_documents():
             enforce_eager=True,
             quantization="awq",
             gpu_memory_utilization=1,
-            max_model_len=14000,
+            max_model_len=12288,
             dtype=torch.float16
         )
         model = models.VLLM(llm)