BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
BEGIN:VEVENT
DTSTAMP:20260730T095926Z
UID:dfad610f-9805-4006-a290-806fac86693b
DTSTART:20260820T140000Z
DTEND:20260820T153000Z
DESCRIPTION:Speakers: Lewis Tunstall (Hugging Face)\n\nAbstract: What does 
 it actually take to train a strong language model? Published research make
 s it look clean: strategic architecture choices\, carefully curated data\,
  enough compute\, and a tidy set of ablations where every decision seems o
 bvious in hindsight. But those reports are written with a fair bit of rosy
  retrospection. They don't capture the 2 a.m. debugging sessions\, the los
 s spikes that appear from nowhere\, or the subtle tensor-parallelism bug t
 hat quietly sabotages a run for days. The reality is messier and more iter
 ative\, and most of the decisions that actually mattered never make it int
 o the final paper. In this lecture I'll walk through the nuts and bolts of
  training an LLM end to end\, from pretraining over trillions of tokens to
  post-training with reinforcement learning. Bio: Lewis is a Machine Learn
 ing Engineer at Hugging Face where he works on applying Transformers to au
 tomate business processes and solve MLOps challenges. Lewis has built ML a
 pplications for startups and enterprises in the domains of NLP\, topologic
 al data analysis\, and time series. In a previous life\, Lewis was a theor
 etical physicist and contributer to open-source projects.\nThis seminar is
  part of the CERN STEAM Academy Seminar Series. \nNetworking cocktail wil
 l follow the seminar. \n\nWith the support of CERN's Next Generation Tri
 ggers Project.\n\n \n\n\nhttps://indico.cern.ch/event/1696055/\n\nZoom: h
 ttps://cern.zoom.us/j/69075966602?pwd=O05Cz0xnV7ay084tmC9VIhlBTlnhS0.1
LOCATION:500/1-001 - Main Auditorium (CERN)
SUMMARY:STEAM Academy Seminar: The Secrets to Training World-Class LLMs
URL;VALUE=URI:https://indico.cern.ch/event/1696055/
END:VEVENT
END:VCALENDAR