Deep Learning VM Images provides containers and virtual machine images to simplify the
configuration of your machine learning (ML) workloads. These images contain the
operating system, the ML frameworks, drivers, and other libraries. We publish
new versions of images regularly to include new patches, security updates, and
features. Each image provided by Deep Learning VM provides support for a
specific minor version of an ML framework.

This allows you time to update and test your code
when moving from one framework version to another. You should always test your
jobs and models thoroughly when switching to a new framework version, regardless
of whether it's a major or minor update.

The Deep Learning VM release notes are a comprehensive log of all the
changes, updates, and new features that are rolled out. They are essential for
anyone using these images to stay informed about the latest developments and
to manage their machine learning environments effectively.
Subscribe to the [Deep Learning VM release notes](https://docs.cloud.google.com/deep-learning-vm/docs/release-notes) page
for announcements about deprecations and new version releases for your images
and frameworks.

See also the [list of supported framework versions](https://docs.cloud.google.com/deep-learning-vm/docs/images#supported-frameworks).

## Shared responsibility

Securing your workloads on Deep Learning VM is a shared responsibility. While
Deep Learning VM regularly publishes new versions of images to address
security vulnerabilities, you are responsible for tasks such as the following:

- Manually upgrading to the latest version.

- Ensuring that you properly configured your services to use the latest version.

For more information, see [Shared responsibility](https://docs.cloud.google.com/deep-learning-vm/docs/shared-responsibility).

## Support policy for framework versions

During the supported period for an ML framework version, we will publish new
image versions regularly. The updates may include the following:

- Patch updates for supported frameworks. For example, if we support
  TensorFlow 2.7, and TensorFlow releases
  2.7.1 to address bugs, we will release a new image version.

- Security updates for supported frameworks.

- Non-breaking updates to other packages and software installed on the image.

- Updates to dependencies that have reached end-of-support. For example, if an
  image has Python 3.7 installed and it reaches the end-of-support date, we
  will release a new image version. If the change in dependency may be a
  breaking change, we will update [List of all available versions](https://docs.cloud.google.com/deep-learning-vm/docs/images#supported-frameworks)
  to indicate the change in the dependency.

Once published, an image version is immutable and does not change. You should
always use the latest image version, as earlier versions may have security
vulnerabilities or other critical bugs.

### Support policy schedule

Deep Learning VM supports images for a specific window of time. It is
a common practice in the industry since components, including some
open source components, have to be managed to ensure security and performance.
For Deep Learning VM, the support policy revolves around two key dates:

- **End-of-patch and support date:** After this date, Deep Learning VM
  will no longer publish new image versions for that specific framework version.
  This means no more patch updates, security fixes, or non-breaking updates.
  Existing resources that have been deployed to Deep Learning VM
  will continue to function, but it's recommended to plan your
  migration to a newer, supported framework version.

  To receive troubleshooting support from Deep Learning VM, you may be
  asked to upgrade to a framework version that is within the supported
  time period.
- **End-of-availability date:** After this date, you can no longer use images
  for this framework version. Services may block the creation of new resources
  using these images, and the images will no longer be available for download.

### Image deprecation

When images reach the **End-of-patch and support date** , they are deprecated.
Deprecation means that these images are removed from public visibility and
it's encouraged to use supported images to help ensure security and performance.
However, if you must use a deprecated image, see [Use an image after
deprecation](https://docs.cloud.google.com/deep-learning-vm/docs/images#use-image-after-deprecation).

## What's next

- Review the [list of supported framework versions](https://docs.cloud.google.com/deep-learning-vm/docs/images#supported-frameworks).