DPO started as a rotation project in his Stanford group, says Inception CEO Stefano Ermon

1 Просмотры

  • Спасибо! Поделитесь с друзьями!

    Вам не понравилось видео. Спасибо за то что поделились своим мнением!

aii
Издатель
Stefano Ermon, longtime Stanford professor and now co-founder and CEO of Inception, on what came out of academic AI research. At the scale his lab could train academically, he says, diffusion models beat GANs while being much more stable, and then the whole thing took over and became Stable Diffusion and Midjourney; in his words, all of that started from ideas developed in academia in his lab. He says he worked on FlashAttention as a co-advisor, another academic idea with huge impact in industry. And DPO, the algorithm now used to align LLMs and diffusion models, started out as a rotation project in his group, developed entirely in academia.

Clip from No Priors with Sarah Guo & Elad Gil. Full episode: https://www.youtube.com/watch?v=N1rjtDs8blY

#ai #dpo #diffusion #research
Категория
AI Картинки
Выполните вход или зарегистрируйтесь для отправки комментария.
Комментариев нет.