Azure Machine Learning: Build, Train, and Deploy ML Models
Azure Machine Learning is a cloud-based service that enables you to accelerate and manage the machine learning lifecycle. Build, train, deploy, and manage machine learning models at scale. It integrates with your favourite tools, including open-source frameworks like TensorFlow, PyTorch, and scikit-learn.
Getting Started
Begin your journey with Azure ML by setting up your workspace and exploring the core components. Our quickstart guides will help you get up and running in minutes.
Quickstart: Create your first Azure ML workspace
Tutorial: Train your first model with Azure ML
Core Concepts
Workspaces
An Azure ML workspace is a top-level resource for Azure Machine Learning. It provides a centralized place to work with all the artifacts you create when you use Azure ML. Workspaces are free and can be created through the Azure portal, Azure CLI, or SDK.
Compute Targets
Compute targets are specific Azure resources where you run your training scripts or host your inference endpoints. Azure ML supports various compute targets, including Azure Machine Learning Compute instances, clusters, managed compute clusters, Kubernetes, and virtual machines.
Example: Creating a managed compute cluster:
from azureml.core import Workspace
from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
ws = Workspace.from_config()
# Choose a name for your CPU cluster
cpu_cluster_name = "cpu-cluster"
try:
cpu_cluster = ComputeTarget(workspace=ws, name=cpu_cluster_name)
print('Found existing compute target')
except ComputeTargetException:
compute_config = AmlCompute.provisioning_configuration(vm_size='STANDARD_DS3_V2',
max_nodes=4)
cpu_cluster = ComputeTarget.create(ws, cpu_cluster_name, compute_config)
cpu_cluster.wait_for_completion(show_output=True)
Datasets
Datasets are a fundamental concept in Azure ML for managing and accessing your data. They enable versioning, lineage tracking, and efficient data access for your training jobs.
Experiments
An experiment is a container for running a training job. It helps you organize your runs, track metrics, and compare different model versions.
Pipelines
Azure ML pipelines allow you to orchestrate complex machine learning workflows, from data preparation to model deployment. This enables reproducibility and automation of your ML processes.
Models
Once you've trained a model, you can register it in your Azure ML workspace. This allows for versioning, tracking, and easy deployment to various endpoints.
Endpoints
Endpoints are the gateways to your deployed models, allowing applications to consume them for real-time predictions or batch scoring.
Tutorials
Explore our comprehensive library of tutorials to learn specific tasks and scenarios:
- Automated ML (AutoML) tutorial
- Model interpretability guide
- Responsible AI toolkit
- Deploying models to Azure Kubernetes Service
SDK & CLI
Leverage the powerful Python SDK and Azure CLI for programmatic control and automation of your Azure ML workflows.
Azure ML Python SDK v2 documentation
MLOps with Azure ML
Implement MLOps best practices for robust, scalable, and repeatable machine learning operations. Azure ML provides tools for CI/CD, monitoring, and model governance.
API Reference
Dive deep into the technical details with the full API reference for the Azure ML SDK and REST API.