AWS AI Blog

AWS Deep Learning AMI Now Includes Apache MXNet 0.11 and TensorFlow 1.3

The AWS Deep Learning Amazon Machine Image (AMI) is designed to help you build stable, secure, and scalable deep learning applications on AWS. The AMI comes pre-installed with popular deep learning frameworks. It has GPU drivers and libraries that let you train sophisticated AI models and scale them in the cloud.

The latest release of the AWS Deep Learning AMI is available for Ubuntu and Amazon Linux platforms. The AMI includes several updates, the most notable of which is MXNet 0.11 with Keras Support. MXNet is now available as a backend choice for users running Keras 1.2. MXNet gives Keras users a simple way to set up multiple GPUs for training and superior performance that scales near-linearly. To read more about the benefits of using MXNet as a backend for Keras, refer to NVIDIA’s blog. The AMI also now runs with latest version of TensorFlow, which is version 1.3.0.

Getting started using the AMI is quick and easy. Follow this step-by-step guide to begin working with deep learning in minutes.

Here is the complete list of deep learning frameworks and GPU drivers on the AMI (both Ubuntu and Amazon Linux versions):

  • MXNet 0.11.0
  • TensorFlow 1.3.0
  • Keras 1.2.2 (DMLC fork with MXNet 0.11 support)
  • Caffe 1.0
  • Caffe2 0.8.0
  • CNTK 2.0
  • Theano 0.9.0
  • Torch (master branch)
  • CUDA Toolkit 8.0
  • cuDNN 5.1
  • NVIDIA Driver 375.66

You can find the AMI for Ubuntu and Amazon Linux on the AWS Marketplace.