AWS Deep Learning Base AMI provides a foundational platform for deep learning on AWS EC2 with NVIDIA CUDA, cuDNN, NCCL, GPU Drivers, Intel MKL-DNN, Docker, NVIDIA-Docker, EFA, and AWS Neuron support. This AMI is suitable for deploying your own custom deep learning environment at scale.
For example, for machine learning developers contributing to open source deep learning framework enhancements, the AWS Deep Learning Base AMI provides a foundation for installing your custom configurations and forked repositories to test out new framework features. You could also be a Machine Learning / AI startup with a highly specialized deep learning setup that needs a foundation to run on a cloud-scale infrastructure.

The AWS Deep Learning Base AMI is provided at no additional charge to Amazon EC2 users.

Below are the core components of AWS Deep Learning Base AMI:

• AWS Deep Learning Tools including AWS Elastic Fabric Adapter(EFA) and AWS Neuron.
• NVIDIA Deep Learning Softwares Including NVIDIA GPU Driver, CUDA Toolkit, cuDNN, NCCL, and Fabric Manager.
• Containerization platforms including Docker, and NVIDIA-Docker for build and run GPU accelerated Docker containers.
• Intel Architecture performance library Intel MKL-DNN.
• A collection of popular tools such as awscli, boto3, numpy, scikit-learn, opencv, pandas, matplotlib, graphviz, jupyter, ipython, and more.