BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:SmartAttica Τraining modules for SMEs - Module 9 "Inference with 
 Transformers & Semantic Search with Sentence Transformers"
DTSTART;VALUE=DATE-TIME:20260622T080000Z
DTEND;VALUE=DATE-TIME:20260622T110000Z
DTSTAMP;VALUE=DATE-TIME:20260815T005320Z
UID:indico-event-209@events.grnet.gr
DESCRIPTION:\n\n\n\nGRNET announces\, in the context of SmartAttica EDIH (
 European Digital Innovation Hub)\, the 9th Module of Τraining modules for
  SMEs  with the subject "Inference with Transformers & Semantic Search wit
 h Sentence Transformers". \n\nDate: June 22nd\, 2026\n\nLocation: Online
  via Zoom \n\nPresentation Languages: Greek\, English\n\nInstructors: Nik
 os Bakas (GRNET)\, Roman Dolgopolyi (GRNET)\n\nDuration: 3 hours\n\nDescri
 ption: This is a hands-on introduction to running open Large Language Mode
 ls and turning text into meaning with embeddings. Participants learn how t
 o load a pre-trained model from the HuggingFace Hub\, generate text with s
 treaming\, and control the tokenizer and chat template. The second half co
 vers Sentence Transformers\, showing how text is mapped into vectors and h
 ow cosine similarity reveals semantic relationships between words and sent
 ences - the foundation of modern search and retrieval systems.\n\nTarget A
 udience: This module is designed for SME developers\, technical leads\, an
 d data scientists who want to incorporate natural language processing (NLP
 ) into their projects. It is ideal for those looking to run models in thei
 r own environment and understand the building blocks behind semantic searc
 h and RAG.\n\nLearning Objectives:\n\n By the end of this module\, partic
 ipants will be able to:\n\n\n	Load and run pre-trained language models fro
 m the HuggingFace Hub using the Transformers library.\n	Generate text with
  both streaming and standard inference\, and use the pipeline abstraction.
 \n	Apply chat templates and manage tokenizers\, padding\, and special toke
 ns correctly.\n	Produce embeddings from text using Sentence Transformers.\
 n	Measure semantic similarity with cosine similarity and interpret the res
 ulting vector space. \n\n\nPrerequisites:\n\nParticipants should have:\n\
 n\n	Basic understanding of Python programming.\n	Familiarity with running 
 code in Jupyter/Colab notebooks.\n	Interest in NLP applications.\n	Some ex
 perience with machine learning will be helpful.\n\n\nIndicative Content:\n
 \n\n	The Transformers Library. Introduction to HuggingFace and the ecosyst
 em for loading and running open models.\n	Model and Tokenizer Setup. Downl
 oading a model from the Hub\, configuring the tokenizer\, and handling pad
 /EOS tokens.\n	Chat Templates and Messages. Structuring system and user me
 ssages and applying the model's chat template.\n	Streaming Inference. Gene
 rating text token-by-token with a streamer for a responsive experience.\n	
 Standard Inference and the Pipeline. Running batch generation and using th
 e high-level pipeline object.\n	Introduction to Embeddings. What embedding
 s are and why similar meanings map to nearby vectors.\n	Sentence Transform
 ers in Practice. Encoding words and sentences into vectors with a compact\
 , fast model.\n	Measuring Similarity. Using cosine similarity to compare t
 exts and visualize a similarity matrix.\n	Summary and Q&A. Key takeaways a
 nd open discussion.\n\n\n \n\nThe project is co-funded by the European Un
 ion. Views and opinions expressed are however those of the author(s) only 
 and do not necessarily reflect those of the European Union or the European
  Commission. Neither the European Union nor the granting authority can be 
 held responsible for them. \n\nhttps://events.grnet.gr/event/209/
LOCATION:
URL:https://events.grnet.gr/event/209/
END:VEVENT
END:VCALENDAR