Skip to content

Files

Latest commit

 

History

History
97 lines (70 loc) · 4.17 KB

resource-curated-environments.md

File metadata and controls

97 lines (70 loc) · 4.17 KB
title titleSuffix description services author ms.author ms.reviewer ms.service ms.subservice ms.topic ms.date
Curated environments
Azure Machine Learning
Learn about Azure Machine Learning curated environments, a set of pre-configured environments that help reduce experiment and deployment preparation times.
machine-learning
ssalgadodev
ssalgado
ssalgado
machine-learning
core
reference
10/21/2021

Azure Machine Learning Curated Environments

This article lists the curated environments with latest framework versions in Azure Machine Learning. Curated environments are provided by Azure Machine Learning and are available in your workspace by default. They are backed by cached Docker images that use the latest version of the Azure Machine Learning SDK, reducing the run preparation cost and allowing for faster deployment time. Use these environments to quickly get started with various machine learning frameworks.

Note

Use the Python SDK, CLI, or Azure Machine Learning studio to get the full list of environments and their dependencies. For more information, see the environments article.

Why should I use curated environments?

  • Reduces training and deployment latency.
  • Improves training and deployment success rate.
  • Avoid unnecessary image builds.
  • Only have required dependencies and access right in the image/container. 

Important

To view more information about curated environment packages and versions, visit the Environments tab in the Azure Machine Learning studio.

Training curated environments

PyTorch

Name: AzureML-pytorch-1.10-ubuntu18.04-py38-cuda11-gpu
Description: An environment for deep learning with PyTorch containing the AzureML Python SDK and other Python packages.

  • GPU: Cuda11
  • OS: Ubuntu18.04
  • PyTorch: 1.10

Other available PyTorch environments:

  • AzureML-pytorch-1.9-ubuntu18.04-py37-cuda11-gpu
  • AzureML-pytorch-1.8-ubuntu18.04-py37-cuda11-gpu
  • AzureML-pytorch-1.7-ubuntu18.04-py37-cuda11-gpu

LightGBM

Name: AzureML-lightgbm-3.2-ubuntu18.04-py37-cpu
Description: An environment for machine learning with Scikit-learn, LightGBM, XGBoost, Dask containing the AzureML Python SDK and other packages.

  • OS: Ubuntu18.04
  • Dask: 2021.6
  • LightGBM: 3.2
  • Scikit-learn: 0.24
  • XGBoost: 1.4

Sklearn

Name: AzureML-sklearn-1.0-ubuntu20.04-py38-cpu Description: An environment for tasks such as regression, clustering, and classification with Scikit-learn. Contains the AzureML Python SDK and other Python packages.

  • OS: Ubuntu20.04
  • Scikit-learn: 1.0

Other available Sklearn environments:

  • AzureML-sklearn-0.24-ubuntu18.04-py37-cpu

TensorFlow

Name: AzureML-tensorflow-2.4-ubuntu18.04-py37-cuda11-gpu
Description: An environment for deep learning with TensorFlow containing the AzureML Python SDK and other Python packages.

  • GPU: Cuda11
  • Horovod: 2.4.1
  • OS: Ubuntu18.04
  • TensorFlow: 2.4

Automated ML (AutoML)

Azure ML pipeline training workflows that use AutoML automatically selects a curated environment based on the compute type and whether DNN is enabled. AutoML provides the following curated environments:

Name Compute Type DNN enabled
AzureML-AutoML CPU No
AzureML-AutoML-DNN CPU Yes
AzureML-AutoML-GPU GPU No
AzureML-AutoML-DNN-GPU GPU Yes

For more information on AutoML and Azure ML pipelines, see use automated ML in an Azure Machine Learning pipeline in Python.

Inference curated environments and prebuilt docker images

[!INCLUDE list-of-inference-prebuilt-docker-images]

Support

Version updates for supported environments, including the base images they reference, are released every two weeks to address vulnerabilities no older than 30 days. Based on usage, some environments may be deprecated (hidden from the product but usable) to support more common machine learning scenarios.