Neuron 任务定义示例

托管设备分配示例

以下示例显示了使用 resourceRequirements 参数请求实例上所有 Neuron 设备的任务定义。该方法仅适用于托管实例。

{
    "family": "ecs-neuron",
    "requiresCompatibilities": ["MANAGED_INSTANCES"],
    "networkMode": "awsvpc",
    "cpu": "8192",
    "memory": "16384",
    "executionRoleArn": "${YOUR_EXECUTION_ROLE}",
    "containerDefinitions": [
        {
            "name": "neuron-inference",
            "image": "763104351884.dkr.ecr.us-east-1.amazonaws.com/huggingface-vllm-inference-neuronx:0.11.0-optimum0.4.5-neuronx-py310-sdk2.26.1-ubuntu22.04",
            "essential": true,
            "command": [
                "--model", "${YOUR_HUGGING_FACE_MODEL_ID}",
                "--port", "8080",
                "--tensor-parallel-size", "2",
                "--allow-non-cached-model"
            ],
            "portMappings": [
                {
                    "containerPort": 8080,
                    "protocol": "tcp"
                }
            ],
            "resourceRequirements": [
                {
                    "type": "NeuronDevice",
                    "value": "ALL"
                }
            ]
        }
    ]
}

在此示例中，容器镜像包括一个针对 AWS Neuron 进行优化的 vLLM 推理服务器。镜像入口点从 HuggingFace 下载一个模型，并针对 Neuron 进行编译，然后在端口 8080 上启动一个兼容 OpenAI 的 API 服务器。将 ${YOUR_HUGGING_FACE_MODEL_ID} 替换为您的 HuggingFace 模型 ID。

手动指定设备示例

以下示例显示了 inf1.xlarge 的一个 Linux 任务定义，即结合使用 EC2 启动类型和 linuxParameters.devices 指定 Neuron 设备路径。

{
    "family": "ecs-neuron",
    "requiresCompatibilities": ["EC2"],
    "placementConstraints": [
        {
            "type": "memberOf",
            "expression": "attribute:ecs.os-type == linux"
        },
        {
            "type": "memberOf",
            "expression": "attribute:ecs.instance-type == inf1.xlarge"
        }
    ],
    "executionRoleArn": "${YOUR_EXECUTION_ROLE}",
    "containerDefinitions": [
        {
            "entryPoint": [
                "/usr/local/bin/entrypoint.sh",
                "--port=8500",
                "--rest_api_port=9000",
                "--model_name=resnet50_neuron",
                "--model_base_path=s3://amzn-s3-demo-bucket/resnet50_neuron/"
            ],
            "portMappings": [
                {
                    "hostPort": 8500,
                    "protocol": "tcp",
                    "containerPort": 8500
                },
                {
                    "hostPort": 8501,
                    "protocol": "tcp",
                    "containerPort": 8501
                },
                {
                    "hostPort": 0,
                    "protocol": "tcp",
                    "containerPort": 80
                }
            ],
            "linuxParameters": {
                "devices": [
                    {
                        "containerPath": "/dev/neuron0",
                        "hostPath": "/dev/neuron0",
                        "permissions": [
                            "read",
                            "write"
                        ]
                    }
                ],
                "capabilities": {
                    "add": [
                        "IPC_LOCK"
                    ]
                }
            },
            "cpu": 0,
            "memoryReservation": 1000,
            "image": "763104351884.dkr.ecr.us-east-1.amazonaws.com/tensorflow-inference-neuron:1.15.4-neuron-py37-ubuntu18.04",
            "essential": true,
            "name": "resnet50"
        }
    ]
}