Attention heads in compact vision transformers each recognize small label subsets with little noise, which the authors exploit for diverse ensembles and low-latency hybrid architectures on CIFAR-100.
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
-
Low-latency vision transformers via large-scale multi-head attention
Attention heads in compact vision transformers each recognize small label subsets with little noise, which the authors exploit for diverse ensembles and low-latency hybrid architectures on CIFAR-100.