``` bash # 指定模型堂驻内存 $ curl http://127.0.0.1:11434/api/generate -d '{"model":"qwen3-coder:latest", "keep_alive": -1}' $ curl http://127.0.0.1:11434/api/generate -d '{"model":"qwen2.5:14b", "keep_alive": -1}' ```