Starting Tuesday, March 24, 2020, NVIDIA GTC Digital is offering courses for you to learn AWS best practices to accomplish your ML goals faster and more easily. Registration is free, so register now. The following sessions are available from AWS:

S22492: Train BERT in One Hour Using Massive Cloud Scale Distributed Deep Learning
Learn how we scaled BERT training near-linearly to 2,048 NVIDIA V100 GPUs at AWS and how you can scale your own training jobs to cloud scale. Learn how to create large training clusters and leverage AWS’s new distributed training framework (which was used to achieve the fastest training time for BERT in the cloud).

  • Aditya Bindal, Senior Product Manager, AWS Deep Engine
  • Indu Thangakrishnan, Software Development Engineer, AWS Deep Engine

S22493: Improve ML Training Performance with Amazon SageMaker Debugger
With Amazon SageMaker Debugger, developers can get complete insights into the training process by automating data capture and analysis from training runs without code changes. We’ll take a closer look at how you can define rules to monitor and analyze tensors and watch for issues in your model.

  • Shashank Prasanna, Senior Advocate, AI/ML
  • Satadal Bhattacharjee, Principal Product Manager, AWS Deep Engine

S22483: From Hours to Minutes: The Journey of Optimizing Mask-RCNN and BERT Using MXNet
Training large deep learning models like Mask R-CNN and BERT takes lots of time and compute resources. Using MXNet, the Amazon Web Services deep learning framework team has been working with NVIDIA to optimize many different areas to cut the training time from hours to minutes.

  • Haibin Lin, Applied Scientist, AWS Deep Engine-Engineering
  • Lin Yuan, Software Development Engineer, Amazon Deep Learning SDK

S21179: Calculating Surface Traversability Using Normal Distribution Transform on NVIDIA TX2 in Amazon Scout
Scout is an autonomous robot from Amazon for delivering packages. Mapping its surroundings in real time to identify traversable space is critical for autonomous driving. We’ll discuss how we implemented a technique called “3D Normal Distribution Transform” (3D NDT) on NVIDIA TX2 to calculate surface traversability and help Scout navigate the neighborhood autonomously and safely.

  • Ka Chen, Senior Graphics Engineer, DEX Robotics

S21290: Visuals as a Service (VaaS): How Amazon and Others Create and Use Photoreal On-Demand Product Visuals with RTX Real-Time Raytracing and the Cloud
Learn how RTX real-time raytracing, combined with GPU cloud computing, is driving a shift in the way leading retailers like Amazon create visuals for their products. We’ll show how Amazon is using this approach and the migenius RealityServer platform to create and provide product visuals throughout their workflows.

  • Paul Arden, CEO, migenius
  • Thomas Dideriksen, Senior Software Developer, Amazon

We hope you find these sessions educational and we will add new sessions to the GTC Digital session catalog over the next few weeks. You can also apply for a free trial to use NVIDA GPU-based Amazon EC2 P3 and G4 instances. Hope you have a good GTC Digital conference!


About the Author

Geoff Murase is a Senior Product Marketing Manager for AWS EC2 accelerated computing instances, helping customers meet their compute needs by providing access to hardware-based compute accelerators such as Graphics Processing Units (GPUs) or Field Programmable Gate Arrays (FPGAs). In his spare time, he enjoys playing basketball and biking with his family.

 

 

 

 

from AWS Machine Learning Blog: https://aws.amazon.com/blogs/machine-learning/aws-delivers-sessions-online-at-nvidia-gtc-digital/