Posted On: Jun 30, 2021

We are excited to announce that AWS now offers new framework-specific Deep Learning AMI (DLAMI). Today we launched framework-specific DLAMIs for PyTorch 1.9.0 and TensorFlow 2.5.0, with support for Amazon Linux 2, Ubuntu 18.04, and Ubuntu 20.04. The AWS DLAMIs provide machine learning practitioners and researchers with the infrastructure and tools to accelerate deep learning in the cloud, at any scale. These optimized new images are up to 60% smaller in size, accelerating the time for machine learning practitioners to launch DLAMIs on AWS. We will continue to release and support DLAMI in the previous format that includes multiple frameworks and operating systems within a single AMI.

In addition, these new framework-specific DLAMIs enable customers to ensure backwards compatibility because each contain a single framework, the version of which is updated with each patch. For example, a DLAMI with TensorFlow (TF) 2.3 will always have the same major and minor version but may receive occasional patch version updates. With the new framework-specific DLAMIs, customers can be confident that the latest version of a framework will maintain compatibility with their applications and deep learning libraries while also staying up to date with framework bug fixes and security patches.

The new framework-specific Deep Learning AMIs and backwards-compatible images are available in all regions where AWS DLAMIs are available. To learn more, please visit our documentation and get started with the AWS Deep Learning AMIs today.