NORDICS20 - Assets

Deep Learning on AWS

Amazon Web Services Resources EMEA

Issue link: https://emea-resources.awscloud.com/i/1242450

Contents of this Issue

Navigation

Page 33 of 50

Amazon Web Services Deep Learning on AWS Page 29 7. Create SLURM Job Script (SLURM is an open source scheduling software for HPC jobs) 8. Submit SLURM Job 9. Obtain Insights In the case of AWS managed services, such as Amazon SageMaker, all of the above steps are automated and triggered with single click in the AWS Management Console, single API call, single CLI command, or event based. Let's dive deeper into each solution to review the components and value of each solution. Fully Managed Solution - Use Amazon SageMaker If you are looking for a solution to scale deep learning across your organization, Amazon SageMaker offers an end-to-end solution to support the different steps involved in a deep learning process. Not only does Amazon SageMaker provide native support in the form of a fully managed service, but it also provides flexibility to customize deep learning stacks to take advantage of most recent innovation in drivers and frameworks. The simplicity and flexibility that Amazon SageMaker offers meets the needs of advanced deep learning engineers and deep learning scientists working at the framework level as well as data scientists and developers contributing to deep learning project with minimal background in deep learning. The following chart shows how different Amazon SageMaker components fit into the deep learning landscape to provide an end-to-end deep learning process. Figure 7: Amazon SageMaker solution for deep learning

Articles in this issue

Links on this page

view archives of NORDICS20 - Assets - Deep Learning on AWS