BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Running and Scaling Performant Deep Learning Models on JSC Superco
 mputers
DTSTART:20260310T120000Z
DTEND:20260312T160000Z
DTSTAMP:20260424T194100Z
UID:indico-event-348@spectra-indico-teszt.hinfra.hu
CONTACT:ai-courses-jsc@fz-juelich.de
DESCRIPTION:Course Content:\nIn this workshop\, we will guide you through 
 the first steps of using the JSC supercomputer machines for your own AI a
 pplication. This workshop should be tailored to your needs – and our tea
 m will guide you through questions like:\n\nHow do I get access to the J
 SC machines?  \nHow do I use the pre-installed\, optimised software? \
 nHow can I run my own code? \nHow can I store data so I can access it 
 fast in training? \nHow can I parallelise my training and use more than
  one GPU? \nHow can I profile AI code to identify bottlenecks and impro
 ve efficiency?  \nHow can I evaluate whether my code can scale beyond 
 a single node?\n\nIn this workshop\, we will try to get your code and your
  workflow running and make your start on the JSC supercomputers as smooth 
 as possible. \nYou will leave with a working setup and a clearer under
 standing of how to iterate efficiently – whether that means speeding up
  a single GPU run\, scaling to multiple GPUs\, or moving to multi-node tra
 ining.\nAfter this course\, you are not only ready to use Helmholtz AI COm
 pute REssources (HAICORE) but you have made your first steps into unlockin
 g compute resources even on the largest scale with a compute time applic
 ation at the Gauss Supercomputing Center.\nThis workshop will be held in
  a small group setting with enough space to address your questions. Pleas
 e give us an indication on what topics you are interested in and we will t
 ry to adjust.\nPrerequisites:\nFor this course\, you should have some init
 ial experience with running Python and machine learning frameworks like Te
 nsorFlow and PyTorch.\nA personal institutional email address (university/
 research institution\, government agency\, organisation\, or company) is r
 equired to register for JSC training courses. If you don't have an institu
 tional email address\, please get in touch with the contact person for thi
 s course.\nLanguage:\nThis course is given in English.\nDuration:\n3 half 
 days\nDates:\n10-12 March 2026\, 13:00-17:00 each day\nVenue:\nOnline\nNum
 ber of Participants:\nMinimum 5\, maximum 25\nInstructors:\nSabrina Benass
 ou\, Ismail Khalfaoui\, Javad Kasravi (JSC)\n\nhttps://spectra-indico-tesz
 t.hinfra.hu/event/348/
LOCATION:Online
URL:https://spectra-indico-teszt.hinfra.hu/event/348/
END:VEVENT
END:VCALENDAR