BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Prilagajanje velikih jezikovnih modelov – praktičen pristop za 
 poslovno in znanstveno rabo | Adapting and fine-tuning LLMs – a hands-on
  approach for business and science
DTSTART:20261120T080000Z
DTEND:20261120T150000Z
DTSTAMP:20261114T050700Z
UID:indico-event-4019@indico.ijs.si
CONTACT:events@slaif.si
DESCRIPTION:Speakers: Matej Klemen (University of Ljubljana\, Faculty of C
 omputer and Information Science)\, Živa Štebljaj (UL FRI)\, Žagar Aleš
  (UL FRI)\, Tjaša Arčon (UL FRI)\, Marko Robnik-Šikonja (UL FRI)\, Dome
 n Vreš (UL FRI)\n\nIzvajalec / Course provider: University of Ljubljana\,
  Faculty of Computer and Information Science (UL FRI)Predavatelji / Instru
 ctors: Marko Robnik-Šikonja\, Matej Klemen\, Aleš Žagar\, Domen Vreš\,
  Tjaša Arčon\, Živa Štebljaj\; (UL FRI)Learning objectives: Get practi
 cal knowledge on how to use large language models (LLMs) via API calls\, t
 heir fine-tuning for specific tasks\, and efficient inference.\nCourse con
 tents: To use large language models (LLMs) in a research or business setti
 ng\, it is necessary to i) to access them on local computers or on GPU ser
 vers via API\, ii) adapt them to specific needs by fine-tuning them.  The
  hands-on workshop will present theoretically and practically\, step-by-st
 ep:\n\nHow LLMs can be used via API on a local or remote server.\nHow to f
 ine-tune encoder-only models such as BERT.\nHow to fine-tune generative mo
 dels such as Llama\, Gemma\, and GaMS using compute-efficient techniques s
 uch as LoRA.\nHow to implement efficient LLM inference.\n\n \nThe work wi
 ll be based on practical datasets and problems from business and research 
 environments. The use cases will cover popular tasks such as sentiment pre
 diction\, information retrieval\, and question answering.\nThe independent
  work consists of an assignment using the dataset of your choice\, prefera
 bly from your business or scientific domain (to be approved by the worksho
 p team). The assignment shall fine-tune a suitable LLM on the prepared dat
 a\, evaluate the results\, and write a report\, presenting the problem\, m
 ethodology\, evaluation\, results\, and conclusions.\nLearning outcomes: P
 ractical knowledge of LLM API use and fine-tuning. Practical problem-solvi
 ng skills with LLMs. Ability to independently solve problems with LLMs.\n\
 nhttps://indico.ijs.si/event/4019/
IMAGE;VALUE=URI:https://indico.ijs.si/event/4019/logo-3560331715.png
LOCATION:TBA (UL Fakulteta za računalništvo in informatiko)
URL:https://indico.ijs.si/event/4019/
END:VEVENT
END:VCALENDAR