ModelArts Service¶

ModelArts is a one-stop development platform for AI developers. With distributed training, automated model building, and model deployment, ModelArts helps AI developers quickly build models and efficiently manage the AI development lifecycle.

Devenv Instance ¶

List Devenv Instances ¶

This interface is used to query devenv instances list. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

for instance in conn.modelartsv1.devenv_instances(de_type="Notebook"):
    print(instance)

Create Devenv Instance ¶

This interface is used to create a devenv instance with parameters. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

attrs = {
    "name": "notebook-d115",
    "flavor": "modelarts.vm.cpu.2u",
    "spec": {
        "storage": {
            "location": {
                "path": "/test-bucket/notebooks/",
            },
            "type": "obs",
        }
    },
    "profile_id": "Multi-Engine 1.0 (python3)-cpu",
}

devenv_instance = conn.modelartsv1.create_devenv_instance(**attrs)
print(devenv_instance)

Get Devenv Instance ¶

This interface is used to get a devenv instance by ID or an instance of class. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

instance_id = "notebook_id"
response = conn.modelartsv1.get_devenv_instance(instance_id)
print(response)

Find Devenv Instance ¶

This interface is used to find a devenv instance by id or name. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "notebook-1234"
response = conn.modelartsv1.find_devenv_instance(
    name_or_id, ignore_missing=False
)
print(response)

Start Devenv Instance ¶

This interface is used to start a devenv instance by id or an instance of class. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

instance_id = "notebook_id"
response = conn.modelartsv1.start_devenv_instance(instance_id)
print(response)

Stop Devenv Instance ¶

This interface is used to stop a devenv instance by id or an instance of class. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

instance_id = "notebook_id"
response = conn.modelartsv1.stop_devenv_instance(instance_id)
print(response)

Delete Devenv Instance ¶

This interface is used to delete a devenv instance by ID or an instance of class. Devenv.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

instance_id = "devenv_instance_id"
conn.modelartsv1.delete_devenv_instance(instance_id, ignore_missing=False)

ModelArts Service ¶

List Services ¶

This interface is used to query modelarts services list. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

for service in conn.modelartsv1.services():
    print(service)

Create Service ¶

This interface is used to create a modelarts service with parameters. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

attrs = {
    "service_name": "mnist",
    "description": "mnist service",
    "infer_type": "real-time",
    "config": [
        {
            "model_id": "xxxmodel-idxxx",
            "weight": "70",
            "specification": "modelarts.vm.cpu.2u",
            "instance_count": 1,
            "envs": {
                "model_name": "mxnet-model-1",
                "load_epoch": "0",
            },
        },
        {
            "model_id": "xxxxxx",
            "weight": "30",
            "specification": "modelarts.vm.cpu.2u",
            "instance_count": 1,
        },
    ],
}

service = conn.modelartsv1.create_service(**attrs)
print(service)

Get Service ¶

This interface is used to get a service by ID or an instance of class. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

service_id = "fb923e36-a239-40f1-ba62-116a50f53f56"
response = conn.modelartsv1.get_service(service_id)
print(response)

Find Service ¶

This interface is used to find a service by Id or name. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "test-service"
response = conn.modelartsv1.find_service(name_or_id, ignore_missing=False)
print(response)

Start Service ¶

This interface is used to start a Service by Id or an instance of class. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "test-service"
service = conn.modelartsv1.find_service(name_or_id)

response = conn.modelartsv1.start_service(service)
print(response)

Stop Service ¶

This interface is used to stop a service by Id or an instance of class. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "test-service"
service = conn.modelartsv1.find_service(name_or_id)

response = conn.modelartsv1.stop_service(service)
print(response)

Delete Service ¶

This interface is used to delete a service by Id or an instance of class. Service.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "test-service"
service = conn.modelartsv1.find_service(name_or_id, ignore_missing=False)

conn.modelartsv1.delete_service(service)

ModelArts Model ¶

List Models ¶

This interface is used to query models list. Model.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

for model in conn.modelartsv1.models():
    print(model)

Create Model ¶

This interface is used to create a model with parameters. Model.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

attrs = {
    "model_name": "mnist",
    "model_version": "1.0.0",
    "source_location": "https://models.obs.xxxx.com/mnist",
    "source_job_id": "55",
    "source_job_version": "V100",
    "model_type": "TensorFlow",
    "runtime": "python2.7",
    "description": "mnist model",
    "execution_code": "https://testmodel.obs.xxxx.com/customize_service.py",
    "input_params": [
        {
            "url": "/v1/xxx/image",
            "protocol": "http",
            "method": "post",
            "param_name": "image_url",
            "param_type": "string",
            "min": 0,
            "max": 9,
            "param_desc": "http://test/test.jpeg",
        }
    ],
    "output_params": [
        {
            "url": "/v1/xxx/image",
            "protocol": "http",
            "method": "post",
            "param_name": "face_location",
            "param_type": "box",
            "param_desc": "face_location param value description",
        }
    ],
    "dependencies": [
        {
            "installer": "pip",
            "packages": [
                {
                    "package_name": "numpy",
                    "package_version": "1.5.0",
                    "restraint": "ATLEAST",
                }
            ],
        }
    ],
    "model_algorithm": "object_detection",
    "model_metrics": '{"f1":0.52381,"recall":0.666667,\
        "precision":0.466667,"accuracy":0.625}',
    "apis": [
        {
            "url": "/v1/xxx/image",
            "protocol": "http",
            "method": "post",
            "input_params": {
                "type": "object",
                "properties": {"image_url": {"type": "string"}},
            },
            "output_params": {
                "type": "object",
                "properties": {"face_location": {"type": "box"}},
            },
        }
    ],
}

model = conn.modelartsv1.create_model(**attrs)
print(model)

Get Model ¶

This interface is used to get a model by ID or an instance of class. Model.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

model_id = "model_id"
response = conn.modelartsv1.get_model(model_id)
print(response)

Find Model ¶

This interface is used to find a model by Id or name. Model.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "mnist-model-1234"
response = conn.modelartsv1.find_model(name_or_id, ignore_missing=False)
print(response)

Delete Model ¶

This interface is used to delete a model by Id or an instance of class. Model.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

name_or_id = "mnist-model-1234"
model = conn.modelartsv1.find_model(name_or_id, ignore_missing=False)

conn.modelartsv1.delete_model(model)

ModelArts Training Job ¶

List Training Jobs ¶

This interface is used to query training jobs list. TrainingJob.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

for training_job in conn.modelartsv1.training_jobs():
    print(training_job)

Create Training Job ¶

This interface is used to create a training job with parameters. TrainingJob.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

attrs = {
    "job_name": "TestModelArtsJob",
    "job_desc": "This is a ModelArts job",
    "workspace_id": "af261af2218841ec960b01ab3cf1a5fa",
    "config": {
        "worker_server_num": 1,
        "app_url": "/usr/app/",
        "boot_file_url": "/usr/app/boot.py",
        "parameter": [
            {
                "label": "learning_rate",
                "value": "0.01",
            },
            {
                "label": "batch_size",
                "value": "32",
            },
        ],
        "dataset_id": "38277e62-9e59-48f4-8d89-c8cf41622c24",
        "dataset_version_id": "2ff0d6ba-c480-45ae-be41-09a8369bfc90",
        "spec_id": 1,
        "engine_id": 1,
        "train_url": "/usr/train/",
        "log_url": "/usr/log/",
    },
}

trainingjob = conn.modelartsv1.create_training_job(**attrs)
print(trainingjob)

Delete Training Job ¶

This interface is used to delete a training job by Id or an instance of class. TrainingJob.

import openstack

openstack.enable_logging(True)
conn = openstack.connect(cloud="otc")

job_id = 123
response = conn.modelartsv1.delete_training_job(job_id)
print(response)

last updated: UTC - commit: