big_vision icon indicating copy to clipboard operation
big_vision copied to clipboard

Published 20 hours ago • verified publisher icon google-research

Mixup Per Example?

Open dibyaghosh opened this issue 1 year ago • 0 comments

Hi! I was wondering why the implementation of mixup uses a single sampled $a$ per batch as opposed to using a different sample $a$ per batch element. Intuitively, it seems that doing this should lead to higher variance in the optimization process.

https://github.com/google-research/big_vision/blob/474dd2ebde37268db4ea44decef14c7c1f6a0258/big_vision/utils.py#L1086

dibyaghosh avatar Feb 06 '24 21:02 dibyaghosh