Update app.py

app.py

@@ -244,7 +244,7 @@ def load_models_and_documents():
             tensor_parallel_size=1,
             enforce_eager=True,
             gpu_memory_utilization=1,
-            max_model_len=10240,
+            max_model_len=8192,
             dtype=torch.float16,
             max_num_seqs=64,
             quantization="gptq"