Knowledge Graph — Coursera Notes › Academic disciplines › Information Technology / Computer Science › Artificial Intelligence › Machine Learning & Data › Azure Machine Learning Service

Azure ML SDK

service · part of Azure Machine Learning Service

A tool for lifecycle support including retraining without service disruption.

The SDK is used with Python and provides classes like Workspace, Datastore, Dataset, Environment, ComputeTarget, ScriptRunConfig, and Experiment from the azureml.core module to programmatically manage Azure ML resources.

To submit a training job, create an environment with conda dependencies (e.g., scikit-learn, pandas, numpy, joblib), register it, define a compute target (e.g., compute instance), and create a ScriptRunConfig specifying the script directory, script name, compute target, and arguments. Then submit the experiment using Experiment.submit(config) and wait for completion with run.wait_for_completion(show_output=True). The output includes logs and execution summary.

from azureml.core import Environment, ComputeTarget, ScriptRunConfig, Experiment
from azureml.core.conda_dependencies import CondaDependencies

env = Environment('myenv')
env.python.conda_dependencies = CondaDependencies.create(conda_packages=['scikit-learn','pandas','numpy','joblib'])
env.register(ws)

compute_target = ComputeTarget(ws, 'your-compute-instance-name')

src = ScriptRunConfig(source_directory='.', script='train.py',
                      arguments=['--data', dataset.as_named_input('trainingdata')],
                      compute_target=compute_target, environment=env)

exp = Experiment(ws, 'fraud-detection')
run = exp.submit(src)
run.wait_for_completion(show_output=True)

Inside Azure ML SDK (1)

• AutoML — An Azure ML SDK feature that automates model and hyperparameter selection for baselines or non-experts.

Connections

• Related to Azure Machine Learning Service
• Alternative to TensorFlow
• Used for Azure ML Service
• Related to Azure Machine Learning Service

Also known as: Azure ML SDK, ML SDK

This is the text view of an interactive 3D knowledge graph — open this page with JavaScript enabled to explore it visually.