Skip to main content

​
Azure Machine Learning Overview

Azure Machine Learning is a cloud service for accelerating and managing the machine learning project lifecycle. Machine learning professionals, data scientists, and engineers can use it to train and deploy models at scale, and to manage MLOps workflows.
Azure Machine Learning provides a unified platform for the complete machine learning lifecycle, from data preparation through model deployment and monitoring.

​
Key Concepts

Azure Machine Learning includes several resources and assets to enable you to perform your machine learning tasks:

​
Resources

Setup or infrastructural resources needed to run a machine learning workflow:

Workspace

Top-level resource providing centralized place to work with all artifacts

Compute

Designated compute resources for training and inference

Datastore

Securely store connection information for data storage

​
Assets

Versioned assets created and registered in your workspace:

Models

Trained machine learning models tracked and versioned

Environments

Encapsulation of software packages and settings

Data

URIs and tables for training and inference

Components

Reusable pipeline steps for ML workflows

​
Azure Machine Learning SDK

The Python SDK v2 provides a programmatic interface to Azure Machine Learning: 
from azure.ai.ml import MLClient
from azure.ai.ml.entities import Workspace
from azure.identity import DefaultAzureCredential

# Connect to your subscription
subscription_id = "<SUBSCRIPTION_ID>"
resource_group = "<RESOURCE_GROUP>"

ml_client = MLClient(
    DefaultAzureCredential(), 
    subscription_id, 
    resource_group
)

​
Azure CLI Extension

The Azure CLI ml extension (v2) enables machine learning operations from the command line: 
az login
az account set --subscription <SUBSCRIPTION_ID>

# Create a workspace
az ml workspace create --file workspace.yml

​
Machine Learning Workflow

1

Prepare Data

Connect to datastores and create data assets for training
2

Train Models

Submit training jobs to compute targets with your training script
3

Register Models

Store and version trained models in the model registry
4

Deploy Models

Create endpoints for real-time or batch inference
5

Monitor

Track model performance and data drift in production

​
Development Environments

Azure Machine Learning supports multiple development tools:
Web-based interface for no-code and code-first experiences
  • Notebooks for interactive development
  • Automated ML for no-code model training
  • Designer for drag-and-drop workflows

​
Storage Format Support

Azure Machine Learning supports three types of storage formats for models:
FormatDescription
custom_modelStandard model format
mlflow_modelMLflow packaged models with metadata
triton_modelNVIDIA Triton inference models

​
Next Steps

Quickstart

Get started with Azure Machine Learning

Training

Learn how to train models

Deployment

Deploy models to production

MLOps

Manage the model lifecycle

Build docs developers (and LLMs) love

Get started for freeTalk to us