Azure Machine Learning compute instance image release notes

  • Article

In this article, learn about Azure Machine Learning compute instance image releases. Azure Machine Learning maintains host operating system images for Azure Machine Learning compute instance and Data Science Virtual Machines. Due to the rapidly evolving needs and package updates, we target to release new images every month.

Azure Machine Learning checks and validates any machine learning packages that might require an upgrade. Updates incorporate the latest OS-related patches from Canonical as the original Linux OS publisher. In addition to patches applied by the original publisher, Azure Machine Learning updates system packages when updates are available. For details on the patching process, see Vulnerability Management.

Main updates provided with each image version are described in the below sections.

September 27, 2024

Image Version : 24.09.23

Release Notes:

azureml-dataprep: 5.2.0

azureml-fsspec: 1.3.1

torch: 2.4.1

cuda-drivers-fabricmanager-535

krb5: 0.6.0

python3-jupyter-core: 4.6.3-3

libzmq5: 4.3.2-2ubuntu1

tensorflow: 2.17.0

python3.8,python3.8-minimal: 3.8.10-0

python3.9,python3.9-minimal: 3.9.5-3

linux-image-5.15.0-1048-gkeop: 5.15.0-1048.55

linux-image-5.15.0-1060-intel-iotg: 5.15.0-1060.66

linux-image-gkeop-5.15: 5.15.0.1053.60

linux-image-intel: 5.15.0.1065.71

linux-image-intel-iotg: 5.15.0.1065.71

libgtk-3-0: 3.24.20-0

libgtk2.0-0: 2.24.32-4

bind9: 1:9.18.28-0

pypy-zipp, python-zipp, python3-zipp: 1.0.0-1

openjdk-11-jdk, openjdk-11-jdk-headless, openjdk-11-jre , openjdk-11-jre-headless, openjdk-11-jre-zero: 11.0.24+8-1ubuntu3

mysql-server-8.0: 8.0.39-0

libssl1.1: 1.1.1f-1ubuntu2.23

snapd: 2.63+20.04ubuntu0.1

curl, libcurl3-gnutls, libcurl3-nss,libcurl4: 7.68.0-1ubuntu2.23

wpasupplicant: 2:2.9-1ubuntu4.4

krb5-admin-server, krb5-kdc, krb5-kdc-ldap, krb5-otp, krb5-pkinit, krb5-user, libgssapi-krb5-2, libgssrpc4, libk5crypto3, libkadm5clnt-mit11, libkadm5srv-mit11, libkdb5-9, libkrad0 libkrb5-3, libkrb5support0: 1.17-6ubuntu4.6

August 2, 2024

Image version: 24.07.01

Release Notes:

Ray: 2.31.0

Nvidia-docker2

Tensorflow: 2.15.0

Pandas: 1.3.5

Libcurl: 8.4.0

Libzmq5: 4.3.2-2ubuntu1

Less: 551-1ubuntu0.3

Libgit2: 0.28.4+dfsg.1-2ubuntu0.1

Klibc-utils: 2.0.7-1ubuntu5.2

Libklibc: 2.0.7-1ubuntu5.2

Libc6: 2.31-0ubuntu9.16

Linux-image-azure: 5.15.0.1045.52

Bind9: 1:9.16.48-0ubuntu0

Binutils: 2.34-6ubuntu1.9

Binutils-multiarch: 2.34-6ubuntu1.9

Libxml2: 2.9.10+dfsg-5ubuntu0

Libuv1: 1.34.2-1ubuntu1.5

Curl: 7.68.0-1ubuntu2.22

Libcurl3-gnutls: 7.68.0-1ubuntu2.22

Libcurl3-nss: 7.68.0-1ubuntu2.22

Libcurl4: 7.68.0-1ubuntu2.22

Util-linux: 2.34-0.1ubuntu9.6

Libglib2.0-0: 2.64.6-1~ubuntu20.04.7

Libglib2.0-bin: 2.64.6-1~ubuntu20.04.7

Gstreamer1.0-plugins-base: 1.16.3-0ubuntu1.3

Xserver-xorg-core: 2:1.20.13-1ubuntu1

Xwayland: 2:1.20.13-1ubuntu1

Libnss3: 2:3.98-0ubuntu0.20.04.2

Accountsservice: 0.6.55-0ubuntu12

Libaccountsservice0: 0.6.55-0ubuntu12

Libssl1.1: 1.1.1f-1ubuntu2.22

Libnode64: 10.19.0~dfsg-3ubuntu1.6

Nodejs: 10.19.0~dfsg-3ubuntu1.6

Libnss3: 2:3.98-0ubuntu0.20.04.2

Libgnutls30: 3.6.13-2ubuntu1.11

Cpio: 2.13+dfsg-2ubuntu0.4

Libtss2-esys0: 2.3.2-1ubuntu0

July 3, 2024

Image Version: 24.06.10 SDK Version: 1.56.0

Issue fixed: Compute Instance 20.04 image build with SDK 1.56.0

Major: Image Version: 24.06.10

SDK （azureml-core): 1.56.0

Python: 3.9

CUDA: 12.2

CUDnn==9.1.1

Nvidia Driver: 535.171.04

PyTorch: 1.13.1

TensorFlow: 2.15.0

Autokeras: 1.0.16

Keras: 2.15.0

Ray: 2.2.0

Docker version: 24.0.9-1

February 16, 2024

Version: 24.01.30

Main changes:

  • Enable Tensorflow in GPU compute to detect GPU device.

Main environment specific updates:

  • N/A

June 30, 2023

Version: 23.06.30

Main changes:

  • Azure Machine Learning SDK to version 1.51.0
  • Purged vulnerable packages
  • Fixed libtinfo error
  • Fixed 'conda command not found' error

Main environment specific updates:

  • tensorflow updated to 2.11.1 in azureml_py38_PT_TF
  • azure-keyvault-keys updated to 4.8.0 in azureml_py38

April 7, 2023

Version: 23.04.07

Main changes:

  • Azure Machine Learning SDK to version 1.49.0
  • Certifi updated to 2022.9.24
  • .NET updated from 3.1 (end-of-life) to 6.0
  • Pyspark update to 3.3.1 (mitigating Log4j 1.2.17 and common-text-1.6 vulnerabilities)
  • Default intellisense to Python 3.10 on the CI
  • Bug fixes and stability improvements

Main environment specific updates:

  • Azureml_py38 environment is now the default.

January 19, 2023

Version: 23.01.19

Main changes:

  • Added new conda environment jupyter-env
  • Moved Jupyter service to new jupyter-env conda environment
  • Azure Machine Learning SDK to version 1.48.0

Main environment specific updates:

  • Added azureml-fsspec package to Azureml_py310_sdkv2
  • CUDA support resolved for azureml_py38CUDA
  • CUDA support resolved for azureml_py38_PT_TF

September 22, 2022

Version 22.09.22

Main changes:

  • .NET Framework to version 3.1.423
  • Azure Cli to version 2.40.0
  • Conda to version 4.14.0
  • Azure Machine Learning SDK to version 1.45.0

Main environment specific updates:

azureml_py38:

  • azureml-core to version 1.45.0
  • tensorflow-gpu to version 2.2.1

August 19, 2022

Version 22.08.19

Main changes:

  • Base OS level image updates.

July 22, 2022

Version 22.07.22

Main changes:

  • Azcopy to version 10.16.0
  • Blob Fuse to version 1.4.4