A vision transformer reduces self-attention complexity from O(n^2) to O(mn) by using m learnable anchor tokens and a two-step Markov transition between anchors and tokens.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AnchorFormer: Differentiable Anchor Attention for Efficient Vision Transformer
A vision transformer reduces self-attention complexity from O(n^2) to O(mn) by using m learnable anchor tokens and a two-step Markov transition between anchors and tokens.