Upgrade local runs to SDK v2

  • Article

Local runs are similar in both V1 and V2. Use the "local" string when setting the compute target in either version.

This article gives a comparison of scenario(s) in SDK v1 and SDK v2.

Submit a local run

  • SDK v1

    from azureml.core import Workspace, Experiment, Environment, ScriptRunConfig

# connect to the workspace
ws = Workspace.from_config()

# define and configure the experiment
experiment = Experiment(workspace=ws, name='day1-experiment-train')
config = ScriptRunConfig(source_directory='./src',
                            script='train.py',
                            compute_target='local')

# set up pytorch environment
env = Environment.from_conda_specification(
    name='pytorch-env',
    file_path='pytorch-env.yml')
config.run_config.environment = env

run = experiment.submit(config)

aml_url = run.get_portal_url()
print(aml_url)

  • SDK v2

    #import required libraries
from azure.ai.ml import MLClient, command
from azure.ai.ml.entities import Environment
from azure.identity import DefaultAzureCredential

#connect to the workspace
ml_client = MLClient.from_config(DefaultAzureCredential())

# set up pytorch environment
env = Environment(
    image='mcr.microsoft.com/azureml/openmpi3.1.2-ubuntu18.04',
    conda_file='pytorch-env.yml',
    name='pytorch-env'
)

# define the command
command_job = command(
    code='./src',
    command='train.py',
    environment=env,
    compute='local',
)

returned_job = ml_client.jobs.create_or_update(command_job)
returned_job

Mapping of key functionality in SDK v1 and SDK v2

Functionality in SDK v1 Rough mapping in SDK v2
experiment.submit MLCLient.jobs.create_or_update

Next steps