AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting

Xiaohan Ding, Tianxiang Hao, Jianchao Tan, Ji Liu, Jungong Han, Yuchen Guo, Guiguang Ding

ICCV 2021 · Architectures and efficiency › Pruning and efficient tuning · approach: Structured network pruning · 205 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
use 1x1 conv compactor (convolutional re-parameterization) instead of channel-wise vector (ablation)ResRep · ImageNet-1K val (ResNet-50)top-1 accuracy (%): 75.57 -> 76.15Improved
raise compactor momentum from 0.9 to 0.99 (ablation)ResRep · ImageNet-1K val (ResNet-50)FLOPs reduction (%): 45.1 -> 54.5Improved
raise compactor momentum from 0.9 to 0.99 (ablation)ResRep · ImageNet-1K val (ResNet-50)top-1 accuracy (%): 75.05 -> 76.15No clear effect
use ResRep instead of pruning then finetuningResNet-50 · ImageNet-1K val (ResNet-50)top-1 accuracy (%): 74.66 -> 76.15Improved
use ResRep instead of a uniformly shrunk network trained from scratchResNet-50 · ImageNet-1K val (ResNet-50)top-1 accuracy (%): 74.39 -> 75.97Improved
finetune the base model for 180 more epochsResNet-50 · ImageNet-1K val (ResNet-50)top-1 accuracy (%): 76.15 -> 76.19No clear effect
add Rep (compactor) without Res to Lasso baseline (ablation)Lasso pruning · CIFAR-10 ResNet-56accuracy at matched FLOPs reductionImproved
add Res (gradient resetting) without Rep to Lasso baseline (ablation)Lasso pruning · CIFAR-10 ResNet-56accuracy at matched FLOPs reductionImproved
combine Rep and Res (full ResRep) (ablation)Rep only or Res only · CIFAR-10 ResNet-56accuracy at matched FLOPs reductionImproved
use ResRep instead of SASL at about 56% FLOPs reductionResNet-50 · ImageNet-1K val (ResNet-50)top-1 accuracy drop (%): 1.00 -> 0.18Improved

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