{
    "inference": {
        "model": [
            {
                "id": "gpt2",
                "instanceType": "ml.g5.12xlarge",
                "numGpu": 4,
                "containerStartupHealthCheckTimeout": 900,
                "quantizationMethod": "gptq"
            },
            {
                "id": "test/gpt2",
                "instanceType": "ml.g5.12xlarge",
                "numGpu": 4,
                "quantizationMethod": "gptq"
            }
        ]
    },
    "training": {}
}