Skip to content
Sign in
  • Browse all toolsFilter by category and pricing
  • Browse categoriesWriting, code, image, audio and more
  • Compare toolsPut up to three side by side
  • Read the blogGuides and comparisons

Azure Machine Learning

Build, train, evaluate, and deploy machine learning models on Azure

Azure Machine Learning is a managed Azure service for building, training, evaluating, and deploying machine learning models. Azure Machine Learning serves data scientists and developers working on predictive models and language model applications. There is no separate service fee, but compute and other Azure services used are billed.

Updated 22 Nov 20266 sections

Azure Machine Learning website screenshot

What is Azure Machine Learning?

Azure Machine Learning is a cloud service for building, training, and deploying machine learning models. Azure Machine Learning gives data scientists and developers a central studio to work with model artifacts across the machine learning lifecycle. The service covers both conventional machine learning tasks, such as classification and regression, and language model development, including fine-tuning and workflow evaluation.

What Azure Machine Learning can do

Azure Machine Learning supports data preparation, model training, evaluation, and deployment. Teams can prepare data on Apache Spark clusters, make features reusable across workspaces, and use interpretability tools to examine model behavior. For language model projects, the model catalog provides access to foundation models, while prompt flow helps developers build and evaluate application workflows.

  • Automated machine learning builds models for classification, regression, vision, and natural language processing.
  • The feature store makes features discoverable and reusable across workspaces.
  • The model catalog supports foundation model discovery, fine-tuning, and deployment.
  • Prompt flow supports the design, evaluation, and deployment of language model workflows.
  • Managed endpoints deploy and score models, log metrics, and support controlled rollouts.

How Azure Machine Learning works

Azure Machine Learning provides workspace resources and a studio for model development. A data scientist can run Jupyter notebooks on a cloud workstation or configure no-code training for tabular data through the studio interface. Developers can use the Python SDK to create production pipelines, and teams can use continuous integration and continuous delivery to automate workflows. Apache Spark data preparation in Azure Machine Learning is interoperable with Microsoft Fabric.

Who Azure Machine Learning is for

Azure Machine Learning is for data scientists and developers who need to build models and put them into production. Azure Machine Learning supports analysts training tabular models without code and developers building language model applications. Teams that manage models across workspaces can use registries to collaborate, and teams that need hybrid machine learning can run compute outside a central environment.

  • Train a no-code classification model on tabular data.
  • Fine-tune a foundation model for a language model application.
  • Set up a reproducible training pipeline through the Python SDK.
  • Deploy a trained model for scoring through a managed endpoint.

Azure Machine Learning pricing and plans

Azure Machine Learning has no separate service fee, but the resources used to train and run models are billable. Compute choices include general-purpose CPUs and specialized GPUs. Storage, secrets management, container registries, and monitoring can also generate separate charges when those Azure services are used. The page lists pay-as-you-go Azure access and an Azure trial of up to 30 days.

Azure Machine Learning strengths and limitations

Azure Machine Learning combines no-code training and Python-based pipelines with managed deployment and model evaluation tools. Azure Machine Learning also provides interpretability capabilities and disparity metrics for teams that need to assess fairness. The cost of a workflow depends on which resources it consumes, and Azure Machine Learning is available in supported Azure regions rather than every region.

  • Training and inference use billable compute.
  • Supporting Azure services are billed separately when used.
  • The service is only available in supported Azure regions.

Azure Machine Learning FAQ

What is Azure Machine Learning?

Azure Machine Learning is a cloud service for building, training, evaluating, and deploying machine learning models. Azure Machine Learning provides a studio where data scientists and developers can work with model artifacts, run notebooks, configure automated training, and build workflows that bring models into production.

Is Azure Machine Learning free?

Azure Machine Learning has no separate service fee, but users pay for compute used during model training and inference. Other Azure services consumed in a workflow, such as Azure Blob Storage, Azure Key Vault, Azure Container Registry, and Azure Application Insights, are billed separately.

Can you train models without code in Azure Machine Learning?

Azure Machine Learning supports no-code automated training for tabular data through its studio interface. Azure Machine Learning provides a tutorial for training a classification model without code and also supports automated model training for regression, vision, and natural language processing.

Can Azure Machine Learning fine-tune and deploy foundation models?

Azure Machine Learning supports fine-tuning and deploying foundation models through its model catalog. The catalog includes models from Microsoft, OpenAI, Hugging Face, Meta, and Cohere. Azure Machine Learning also provides prompt flow for designing, evaluating, and deploying language model workflows.

How do you deploy a model with Azure Machine Learning?

Azure Machine Learning provides managed endpoints for deploying models and running model scoring. Managed endpoints can log metrics and support controlled rollouts. Azure Machine Learning also supports reproducible machine learning pipelines and workflow automation through continuous integration and continuous delivery.

From the community

No reviews yet

Be the first to review this tool. What you write helps the next person choose.

More tools like Azure Machine Learning

Amazon SageMaker

aws.amazon.com

Build, train, and deploy AI models and run SQL analytics with unified access to enterprise data.

Paid

Hugging Face

www.huggingface.co

Discover, host and collaborate on machine learning models, datasets and applications.

Freemium

Lightning AI

www.lightning.ai

Build, train and deploy specialized AI models with cloud GPUs, development Studios and enterprise infrastructure.

Freemium

TensorFlow

www.tensorflow.org

Build, train, evaluate, and deploy machine learning models across browsers, servers, and edge devices.

Freemium