How can I quickly use inference on an AGX orin 64GB server? Is the environment configuration supported? What should I pay attention to?