update

examples/deploy/agent/server.sh

@@ -1,4 +1,7 @@
 CUDA_VISIBLE_DEVICES=0 swift deploy \
     --model Qwen/Qwen2.5-7B-Instruct \
     --infer_backend vllm \
+    --gpu_memory_utilization 0.9 \
+    --max_model_len 8192 \
+    --max_new_tokens 2048 \
     --served_model_name Qwen2.5-7B-Instruct