Pith. sign in

Do vision trans- formers see like convolutional neural networks?, 2022

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.CV 1

years

2024 1

verdicts

CONDITIONAL 1

roles

background 1

polarities

support 1

representative citing papers

CNNtention: Can CNNs do better with Attention?

cs.CV · 2024-12-16 · conditional · novelty 3.0

Adding attention blocks between ResNet-20 feature extractors gives small accuracy improvements on CIFAR-10 and MNIST, but the gains are within a range that needs error bars to trust.

citing papers explorer

Showing 1 of 1 citing paper.

  • CNNtention: Can CNNs do better with Attention? cs.CV · 2024-12-16 · conditional · none · ref 7

    Adding attention blocks between ResNet-20 feature extractors gives small accuracy improvements on CIFAR-10 and MNIST, but the gains are within a range that needs error bars to trust.