Machine learning for Python apps on Azure

The following articles help you get started with Azure Machine Learning. Azure Machine Learning v2 REST APIs, Azure CLI extension, and Python SDK are designed to streamline the entire machine learning lifecycle and accelerate production workflows. The links in this article target v2, which is recommended if you're starting a new machine learning project.

Getting started

In Azure Machine Learning, the workspace is the main resource that organizes and manages everything you create, such as datasets, models, and experiments.

• Quickstart: Get started with Azure Machine Learning
• Manage Azure Machine Learning workspaces in the portal or with the Python SDK (v2)
• Run Jupyter notebooks in your workspace
• Tutorial: Model development on a cloud workstation

Deploy models

Deploy models for low-latency, real-time machine learning predictions.

• Tutorial: Designer - deploy a machine learning model
• Deploy and score a machine learning model by using an online endpoint

Automated machine learning

Automated ML (AutoML) refers to the process of streamlining machine learning model development by automating its repetitive and time-consuming tasks.

• Train a regression model with AutoML and Python (SDK v1)
• Set up AutoML training for tabular data with the Azure Machine Learning CLI and Python SDK (v2)

Data access

With Azure Machine Learning, you can import data from your local computer or connect to existing cloud storage services.

• Create and manage data assets
• Tutorial: Upload, access and explore your data in Azure Machine Learning
• Access data in a job

Machine learning pipelines

Use machine learning pipelines to build workflows that connect different stages of the ML process.

• Use Azure Pipelines with Azure Machine Learning
• Create and run machine learning pipelines using components with the Azure Machine Learning SDK v2
• Tutorial: Create production ML pipelines with Python SDK v2 in a Jupyter notebook