-
|
Sorry, but who can tell me why set different seed? |
Beta Was this translation helpful? Give feedback.
Answered by
rwightman
Aug 26, 2022
Replies: 1 comment
-
|
@Fei-Wang so that each rank augments data differently for train pipeline ... except where you need them to be the same (ie distributed sampler), where they use their own random gen... |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
Fei-Wang
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
@Fei-Wang so that each rank augments data differently for train pipeline ... except where you need them to be the same (ie distributed sampler), where they use their own random gen...