BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal//recurring_events_ical//2.0//EN
BEGIN:VEVENT
UID:640a97b5-0039-4da6-9664-88faf741ceff@support.access-ci.org
DTSTAMP:20251121T123849Z
DTSTART:20260414T180000Z
DTEND:20260414T190000Z
SUMMARY:Fine Tuning Large Language Models (LLMs) with Domain Specific Datas
 ets
DESCRIPTION:SummaryLarge language models (LLMs) are trained on massive, pub
 licly available text datasets comprising trillions of tokens, enabling the
 m to excel at general language tasks like next-token prediction. However, 
 LLMs often struggle with domain-specific prompts, exhibiting reduced accur
 acy or generating inaccurate information (hallucinations). This is because
  they lack sufficient subject matter expertise. Two primary approaches exi
 st to address this limitation for augmenting LLMs knowledge: Retrieval-Aug
 mented Generation (RAG) and fine-tuning. This presentation focuses on fine
 -tuning smaller LLMs with domain-specific instruct datasets using the LoRA
  (Low-Rank Adaptation) technique on Gaudi hardware. We will leverage publi
 cly available LLMs and datasets from the Hugging Face Hub for this demonst
 ration. Though it is possible to fine tune LLMs with plain text data – s
 ourced from documents, articles, and other materials.InstructorMadhusudan 
 Gujral is currently a bioinformatics lead at SDSC. His background is in st
 ructural biology, but he transitioned to the field of informatics over 20 
 years ago. He began by developing a client-based laboratory information sy
 stem (LIM) for a distributed biological project with users across the US. 
 This was followed by a large project creating complex pipelines for metage
 nomics research. He then spent a decade processing and analyzing whole gen
 ome sequencing (WGS) data from thousands of samples collected from patient
 s with psychiatric disorders. For the past two years, he has focused on le
 arning and benchmarking fine-tuning large language models (LLMs) on Gaudi 
 hardware.See a complete list of SDSC's upcoming training and events here.
URL:https://support.access-ci.org/events/8717
END:VEVENT
END:VCALENDAR