BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160551Z
LOCATION:Track 4
DTSTART;TZID=America/New_York:20201117T130000
DTEND;TZID=America/New_York:20201117T143000
UID:submissions.supercomputing.org_SC20_sess176@linklings.com
SUMMARY:Memory Efficient Deep Learning
DESCRIPTION:Paper\n\nScaling Distributed Deep Learning Workloads beyond th
 e Memory Capacity with KARMA\n\nWahib, Zhang, Nguyen, Drozd, Domke...\n\nT
 he dedicated memory of hardware accelerators can be insufficient to store 
 all weights and/or intermediate states of large deep learning models. Alth
 ough model parallelism is a viable approach to lessen the memory pressure 
 issue, significant modification of the source code and considerations for 
 alg...\n\n---------------------\nZeRO: Memory Optimizations Toward Trainin
 g Trillion Parameter Models\n\nRajbhandari, Rasley, Ruwase, He\n\nLarge de
 ep learning models offer significant accuracy gains, but training billions
  of parameters is challenging. Existing solutions exhibit fundamental limi
 tations fitting these models into limited device memory, while remaining e
 fficient.  Our solution uses ZeroRedundancy Optimizer (ZeRO) to optimi...\
 n\n---------------------\nKraken: Memory-Efficient Continual Learning for 
 Large-Scale Real-Time Recommendation\n\nXie, Ren, Lu, Yang, Xu...\n\nModer
 n recommendation systems in industry often use deep learning (DL) models t
 hat achieve better model accuracy with more data and model parameters. Cur
 rent open-source DL frameworks, however, such as TensorFlow and PyTorch, s
 how relatively low scalability on training recommendation models with ter.
 ..\n\n\nTag: Machine Learning, Deep Learning and Artificial Intelligence, 
 Memory Optimization, Scalable Computing\n\nRegistration Category: Tech Pro
 gram Reg Pass
END:VEVENT
END:VCALENDAR