BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal//recurring_events_ical//2.0//EN
BEGIN:VEVENT
UID:45b82e73-c18d-4c2d-a97a-ebb4888f86d3@support.access-ci.org
DTSTAMP:20240219T100739Z
DTSTART:20240228T160000Z
DTEND:20240229T000000Z
SUMMARY:Model Parallelism: Building and Deploying Large Neural Networks (NV
 IDIA Deep Learning Institute)
DESCRIPTION:Large language models (LLMs) and deep neural networks (DNNs), w
 hether applied to natural language processing (e.g., GPT-3), computer visi
 on (e.g., huge Vision Transformers), or speech AI (e.g., Wave2Vec 2), have
  certain properties that set them apart from their smaller counterparts. A
 s LLMs and DNNs become larger and are trained on progressively larger data
 sets, they can adapt to new tasks with just a handful of training examples
 , accelerating the route toward general artificial intelligence. Training 
 models that contain tens to hundreds of billions of parameters on vast dat
 asets isn’t trivial and requires a unique combination of AI, high-perfor
 mance computing (HPC), and systems knowledge. The goal of this course is t
 o demonstrate how to train the largest of neural networks and deploy them 
 to production.In this course, students will learn:- Scale training and dep
 loyment of LLMs and neural networks across multiple nodes- Use techniques 
 such as activation checkpointing, gradient accumulation, and various forms
  of model parallelism to overcome the challenges associated with large-mod
 el memory footprint.- Capture and understand training performance characte
 ristics to optimize model architecture.- Deploy very large multi-GPU, mult
 i-node models to production using NVIDIA Triton™ Inference Server.Upon s
 uccessful completion of the assessment, participants will receive an NVIDI
 A DLI certificate to recognize their subject matter competency and support
  professional career growth.
URL:https://support.access-ci.org/events/7358
END:VEVENT
END:VCALENDAR