~/talk

GPT or how I learned to stop worrying and love the generative AI

on YouTube 28m
Talkat PyCon Italia 2023 · 2023-05-25..05-28· [large-language-models, machine-learning]

speaker

name:Tommaso Radicioni
links:none on file

resources

video:https://youtu.be/ByborJli0bY
slides:none on file

summary

Shows how to load and fine‑tune GPT-2 with the transformers library, using a tiny lyrics dataset (10 examples) that fine‑tunes in about five minutes on a laptop GPU, and how decoding choices change outputs: greedy, beam search, sampling (temperature/top‑K/top‑P) versus recommending contrastive search to reduce repetition. Demonstrates a working demo chatbot using a Streamlit/chain setup, mentions NanoGPT for training medium models (claim: ~4 days on a GPU) and uses perplexity as a validation metric.

AI-written summary · about

chapters (6)

  1. 0:00Intro
  2. 0:40What is generative AI
  3. 9:00How to use GPT
  4. 10:25Output
  5. 16:18Finetuning
  6. 21:25Questions

related talks

~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.