Skip to content

Releases: Azure/azureml-sdk-for-r

CRAN Release

30 Sep 03:53
Compare
Choose a tag to compare

New features

  • Add foreach backend for distributed training and batch inferencing. (#215)
  • Add R Section in Environment Definition.
  • Add Azure Data Lake Gen2 Datastore support as an experimental feature.(#279)
  • Add github_package() and cran_package() constructors to specify packages and
    package versions in Estimator and Environment Definition.(#310)
  • Expose query_timeout parameter for create_tabular_dataset_from_sql_query().(#308)
  • Add data_path() so that Dataset constructors can take in DataPath objects. (#271)
  • Add dataset_consumption_config(). (#272)
  • Add support for ResourceConfiguration and registering models from run.(#300)
  • Expose cluster_purpose param for create_aks_compute(). (#276)
  • Add interactive_login_authentication() and service_principal_authentication. (#263) (#241)
  • Deprecate live run widget.

Bug fixes

  • Fix issues with Dataset creation and usage.
  • Fix Interactive Authentication.

Documentation

  • Add "Troubleshooting" article.
  • Add "Deploying models" vignette.
  • Add sample for batch inferencing with foreach backend.
  • Make all vignettes discoverable via CRAN. (#320)

CRAN Release

06 Feb 00:42
d1c1cc8
Compare
Choose a tag to compare
CRAN Release Pre-release
Pre-release

New features

  • Methods for creating and managing Azure ML Datasets.
  • Update create_workspace() to use sku parameter.
  • Expose file_name parameter to load_workspace_from_config().
  • v2 of the Azure ML run details widget in RStudio Viewer pane.

Bug fixes

  • Fix installation issue introduced by latest reticulate 1.14 release.
  • Fix default CRAN CDN.
  • Remove dependency on DAAG package in train-and-deploy-to-aci vignette.

CRAN Release

18 Nov 18:28
e69eea0
Compare
Choose a tag to compare
CRAN Release Pre-release
Pre-release

Initial CRAN release

Initial features

  • Methods for creating and managing Azure Machine Learning (Azure ML) Workspaces.
  • Methods for registering and managing Azure ML Datastores.
  • Methods for managing secrets in the Azure Key Vault associated with a workspace.
  • Methods for creating and managing Azure Machine Learning Compute (AmlCompute).
  • Methods for creating, managing, and attaching Azure Kubernetes Service clusters as compute targets (AksCompute).
  • Methods for creating and managing Azure ML Environments for training and deployment.
  • Methods for creating and submitting Azure ML Experiments.
  • Methods for configuring and managing Azure ML Runs.
  • Methods for logging metrics to Azure ML during runs.
  • Methods for configuring and managing Azure ML hyperparameter tuning runs.
  • Methods for model registration and management to Azure ML.
  • Methods for deploying models as local webservices for testing.
  • Methods for deploying models as webservices on Azure Container Instances (ACI) and Azure Kubernetes Service (AKS).
  • view_run_details() for invoking remote Azure ML run details in RStudio Viewer pane.

Documentation

  • Vignettes for installation, TensorFlow training on AmlCompute, hyperparameter tuning a Keras model with Azure ML's HyperDrive service, and production deloying a model as a webservice to AKS.
  • Additional code samples for setting up workspaces, training on AmlCompute, and deploying a model as a webservice to ACI.

v0.5.6

29 Oct 22:04
51f3b1c
Compare
Choose a tag to compare
v0.5.6 Pre-release
Pre-release

Release for Ignite

v0.5.2

16 Oct 02:23
Compare
Choose a tag to compare
v0.5.2 Pre-release
Pre-release
Update R SDK docs via Build 20191015.4 [skip ci]

v0.4.0

01 Oct 23:25
996da7c
Compare
Choose a tag to compare
v0.4.0 Pre-release
Pre-release

Initial release for AzureML R SDK. R users are able to train and deploy their models using the Azure ML R SDK, while leveraging additional functionality including run tracking and hyperparameter tuning.