TokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers

[LukeWood]

https://arxiv.org/pdf/2207.08409v1.pdf

[innat]

Here is another augmentation strategy for ViT model. TransMix: Attend to Mix for Vision Transformers

[bhack]

RecursiveMix: Mixed Learning with History

https://arxiv.org/abs/2203.06844 https://github.com/implus/RecursiveMix-pytorch

[LukeWood]

Closing issues out that we don't intend to pursue anytime soon.