Thread

Training Text-to-Image Models 3.6× Faster

56 points · 10 comments · schopra909

schopra909 · 1789592268 · focus · HN ↗
Author here, feel free to drop questions below. Will try to answer to best of my ability!
1. popalchemist · 1789602225 · focus · HN ↗
  Does 3.6x faster also = 3.6x cheaper?
  1. schopra909 · 1789602599 · focus · HN ↗
    Yes, 3.6x cheaper when it comes to training!
    When it comes to inference it should also be cheaper (since we’ve cut attention sequence length by 4x), but I don’t have a hard number for you how much cheaper for inference

Unofficial Hacker News client; not affiliated with Y Combinator.