What are AWS Deep Learning Containers?
AWS Deep Learning Containers are prepackaged and fully tested Docker images that, according to AWS, let you deploy deep learning environments in minutes. Instead of packing frameworks, drivers and dependencies into an image yourself, you use ready-made images maintained by AWS.
The images are performance-optimized for the respective frameworks and can be used on Amazon EC2, Amazon ECS and Amazon EKS as well as in Amazon SageMaker. Training and inference workloads therefore share the same reproducible base.
Core capabilities
- Ready-made Docker images: Prepackaged and fully tested, with no build process of your own
- Framework coverage: PyTorch, TensorFlow and Apache MXNet
- Hardware optimization: Variants for AWS Trainium, AWS Inferentia, AWS Graviton and NVIDIA GPUs
- Service integration: Usable on Amazon EC2, Amazon ECS, Amazon EKS and in Amazon SageMaker
- Documented image list: Available images and release notes in the GitHub repository aws/deep-learning-containers
Typical use cases
Training on Kubernetes: Amazon EKS clusters run training jobs in the provided images, so individual teams do not have to maintain their own base images.
Inference endpoints: Amazon SageMaker uses the same images for inference, keeping training and production environments aligned.
Domain workloads: AWS names autonomous vehicle deployment, natural language processing and healthcare data analysis as examples.
Benefits
- No need to build and test framework images yourself
- Consistent environment from development through production
- Optimized variants for AWS accelerators and NVIDIA GPUs
- Traceable versioning through the documented image list
Integration with innFactory
As an AWS Reseller, innFactory supports you with AWS Deep Learning Containers: selecting suitable images, building training and inference pipelines on EKS or SageMaker, and integrating them into existing CI/CD and registry processes.
Typical Use Cases
Technical Specifications
Frequently Asked Questions
What are AWS Deep Learning Containers?
AWS Deep Learning Containers are prepackaged and fully tested Docker images that, according to AWS, let you deploy deep learning environments in minutes. The images are performance-optimized for the respective frameworks.
Which frameworks are supported?
AWS names PyTorch, TensorFlow and Apache MXNet as supported frameworks. The available image variants and versions are documented in the GitHub repository aws/deep-learning-containers.
Which hardware do the containers target?
AWS states the images are built for AWS Trainium, AWS Inferentia, AWS Graviton and NVIDIA GPUs, and are provided on the basis of Ubuntu Linux.
Which AWS services can they be used with?
AWS names Amazon EC2, Amazon ECS and Amazon EKS as well as integration with Amazon SageMaker. The same images are therefore used for training and inference across different execution environments.
Where can I find the available images?
AWS points to the GitHub repository aws/deep-learning-containers. It lists the available images with framework version, processor type and operating system, along with the release notes.
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of AWS (official documentation). This page does not represent an offer by AWS.