AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Pruning Self-Attentions Into Convolutional Layers in Single Path

Haoyu He, Jianfei Cai, Jing Liu, Zizheng Pan, Jing Zhang, Dacheng Tao, Bohan Zhuang

TPAMI 2024 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 72 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
single-path instead of multi-path search (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 70.5 -> 70.7Improved
single-path instead of multi-path search (ablation)SPViT-DeiT-S · ImageNet-1k valTop-1 accuracy: 78.0 -> 78.3Improved
add MSA pruning to FFN-only pruning (ablation)SPViT-DeiT-B · ImageNet-1k valTop-1 accuracy: 81.3 -> 81.5Improved
add hard-label distillation with RegNetY-16GF teacher (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 70.7 -> 73.2Improved
add hard-label distillation (ablation)SPViT-DeiT-B · ImageNet-1k valTop-1 accuracy: 81.5 -> 82.4Improved
use uncompressed DeiT-Ti as teacher (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 70.7 -> 71.1No clear effect
use CNN teacher RegNetY-1.6GF instead of DeiT-Ti teacher (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 71.1 -> 72.8Improved
use higher capacity CNN teacher RegNetY-16GF (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 72.8 -> 73.2Improved
CNN teacher RegNetY-1.6GF instead of DeiT-Ti teacher (ablation)UVC-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 71.3 -> 70.5Hurt
search MSA-to-conv choice instead of random sampling (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 70.2 -> 70.7Improved
search FFN ratios instead of random sampling (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 70.4 -> 70.6Improved
prune attention into convolution instead of random head pruning (ablation)SPViT-DeiT-Ti · ImageNet-1k valTop-1 accuracy: 69.6 -> 70.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy