Exact Linear Attention uses exact kernel decompositions with non-negativity constraints to achieve linear-complexity attention in Transformers, plus new modules for memory and MoE routing.
YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Exact Linear Attention
Exact Linear Attention uses exact kernel decompositions with non-negativity constraints to achieve linear-complexity attention in Transformers, plus new modules for memory and MoE routing.