Posted On: Jun 26, 2019
AWS Deep Learning (DL) Containers now come with libraries and packages required for model training and inference in HAQM SageMaker. HAQM SageMaker is a fully managed service that enables you to build, train, and deploy machine learning models at scale. You can now use DL Containers seamlessly across HAQM SageMaker, HAQM Elastic Container Service for Kubernetes (HAQM EKS), self-managed Kubernetes on HAQM EC2, and HAQM Elastic Container Service (HAQM ECS). This release of DL Containers also updates Apache MXNet images to 1.4.1 with support for CUDA 10.0.
AWS DL Containers are continually updated with the latest deep learning frameworks and libraries. DL Containers provide optimized and validated Docker images that allow developers to easily setup custom machine learning environments for training and inference on HAQM SageMaker, HAQM EC2, HAQM ECS, and HAQM EKS. Docker images are available for TensorFlow and Apache MXNet on CPU and GPU hardware for training and inference. You can get DL Containers through HAQM Elastic Container Registry (HAQM ECR) and AWS Marketplace at no cost--you pay only for the resources that you use.
For more information, see AWS Deep Learning Containers and dockerfiles for DL Containers. To get started with AWS DL Containers, check out the documentation and the tutorial.