11-785 Spring 2023 - Recitation 11 - Transformers

preview_player
Показать описание
Resources:

Errata:
00:38 At the beginning, I ask students to refer to lectures 16-17, when it should be 18-19. Please refer to lectures 18 and 19 by Bhiksha and Abu.
01:16:33 - It should be a lower triangular matrix of ones (not left? My brain is muddled, sorry lol). It is not an upper triangular matrix because we use a comparison of mask == 0 in the attention block. It would be upper triangular if that was inverted.
Рекомендации по теме
Комментарии
Автор

very poor quality video nothing is visible, so 1st dislike from me

anig