AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Isomorphic Pruning for Vision Models

Gongfan Fang, Xinyin Ma, Michael Bi Mi, Xinchao Wang

ECCV 2024 · Architectures and efficiency › Attention and transformers · approach: Vision transformer pruning · 31 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
prune DeiT-S-size model from DeiT-B instead of training DeiT-T from scratch (isomorphic pruning)DeiT-T · ImageNet-1K valtop-1 accuracy (%): 74.52 -> 77.50Improved
isomorphic pruning ConvNext-B to ConvNext-T sizeConvNext-T · ImageNet-1K valtop-1 accuracy (%): 82.06 -> 82.19Improved
group sub-structures by isomorphism and rank within groups instead of global ranking (ablation)Taylor, Hessian, L1 criteria on ResNet-50 · ImageNet-1K val without finetuningaccuracy and lossImproved
use isomorphic pruning on ResNet-101 for 4.5G and 3.8G budgetsResNet-101 · ImageNet-1K valtop-1 accuracy drop: +0.16 and +0.05Improved
isomorphic pruning on lightweight MobileNet-v2 to 0.15G MACsMobileNet-v2 · ImageNet-1K valtop-1 accuracy drop: 71.88 -> 68.91Mixed
global ranking without grouping on DeiT-B at 50% pruning (ablation)naive global pruning · DeiT-B importance distributionpruning balanceHurt
non-uniform pruned DeiT-S instead of uniform DeiT-SDeiT-S · peak memorypeak memory (MB): 1363 -> 1547Hurt
halve MACs of ResNet-50 by pruningResNet-50 · GPU latencyspeedup: 1.33x speedupMixed
transfer pruned DeiT-S to CIFAR-100DeiT-S · CIFAR-100accuracy (%): 87.07 -> 90.09Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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