GTC 2020: Productionizing GPU Accelerated IoT Workloads at the Edge
After clicking “Watch Now” you will be prompted to login or join.
Click “Watch Now” to login or join the NVIDIA Developer Program.
Productionizing GPU Accelerated IoT Workloads at the Edge
Paul DeCarlo, Microsoft | Ian Davis, Microsoft | Jasmine Greenaway, Microsoft
In this session, we will cover a variety of techniques, tools, and services to assist in the production of Computer Vision Solutions deployed in Edge environments that run on NVIDIA Jetson hardware. Topics will cover: Usage of Azure DevOps tooling and workflows to produce deployable container-based GPU-Accelerated modules for use with Azure IoT Edge The Azure IoT Central - Software as a Service offering for fleet management and runtime-configuration of IoT Edge devices Deepstream compatible custom object detection models built with the Azure Custom Vision AI service offering Jetson-containers tooling for producing CUDA compatible container images and flashable CUDA-capable base OS images targeted to NVIDIA Jetson hardware. The presentation will consist of live demonstrations of these topics and include resources for reproduction. At the end of the session, you will leave understanding how to develop production grade GPU-accelerated container workloads backed by deployment, management, and configuration capabilities in Microsoft Azure.