Tag
vision transformers
2 articles

Research/May 7
Outlier Tokens in DiTs, and How DSR Fixes Them
A new paper shows outlier tokens affect both RAE encoders and DiT denoisers, and proposes Dual-Stage Registers to reduce artifacts.

Research/Apr 3
Steerable ViT Features for Text-Guided Vision
A new vision representation lets text steer ViT features toward specific objects without giving up generic visual utility.