BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//programme.europython.eu//europython-2026//talk//899VGG
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-europython-2026-899VGG@programme.europython.eu
DTSTART;TZID=CET:20260717T101000
DTEND;TZID=CET:20260717T105500
DESCRIPTION:Deep learning is often taught through large frameworks and larg
 e models\, which is great for getting real projects out of the door\, but 
 not always great for learning. This talk is about a different practice: bu
 ilding tiny\, runnable versions of various modern architectures with minim
 al dependencies (mostly Python and NumPy) to learn about the ideas through
  application.\n\nWe’ll get our feet wet by building a small Transformer 
 end-to-end and learn about the model architecture that started the craze. 
 Then we  switch perspectives\, and learn about other architectures\, alway
 s staying small and nimble\, focusing on applying the math and breathing l
 ife into formulas. We will look look at multi-scale modelling (in a simpli
 fied version of Renormalizing Generative Models)\, State Spaces\, and othe
 r scary concepts\, until they are not scary at all anymore.\n\nYou’ll le
 ave with a model for turning papers into little prototypes that stay true 
 to ideas and the starting point for your own little lab to build models yo
 urself.\n\nPrerequisites: a basic understanding of NumPy and a willingness
  to look at Greek letters. No deep learning framework knowledge required.
DTSTAMP:20260624T064846Z
LOCATION:S2
SUMMARY:AI Architecture Katas: Learning by Building Small Models in Plain P
 ython - hellerve
URL:https://programme.europython.eu/europython-2026/talk/899VGG/
END:VEVENT
END:VCALENDAR