AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

RED++ : Data-Free Pruning of Deep Neural Networks via Input Splitting and Output Merging

Edouard Yvinec, Arnaud Dapogny, Matthieu Cord, Kévin Bailly

TPAMI 2022 · Architectures and efficiency › Pruning and efficient tuning · approach: Structured network pruning · 27 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
add input splitting step (hashing + merge alpha*) (ablation)RED++ · Cifar10 ResNet 56pruning ratio (% params removed): 77.68 -> 88.81Improved
add input splitting step (hashing + merge alpha*) (ablation)RED++ · Cifar10 ResNet 20pruning ratio (% params removed): 41.03 -> 67.48Improved
add input splitting step (hashing + merge alpha*) (ablation)RED++ · ImageNet ResNet 50pruning ratio (% params removed): 0.29 -> 44.25Improved
add weight hashing to merge alpha* and split (ablation)RED++ · Cifar10 ResNet 56pruning ratio (% params removed): 61.19 -> 88.81Improved
add weight hashing to merge alpha* and split (ablation)RED++ · ImageNet ResNet 50pruning ratio (% params removed): 0.09 -> 44.25Improved
use relaxed merging alpha* instead of alpha=0 (with hashing and split) (ablation)RED++ · Cifar10 ResNet 56pruning ratio (% params removed): 85.89 -> 88.81Improved
use relaxed merging alpha* instead of alpha=0 (with hashing and split) (ablation)RED++ · Cifar10 ResNet 20pruning ratio (% params removed): 65.26 -> 67.48Improved
merge alpha=0 without hashing (ablation)RED++ · Cifar10 and ImageNet networkspruning ratio (% params removed): 0.00No clear effect
use adaptive hashing instead of uniform int8 hashing (ablation)RED++ · Cifar10 ResNet 56% accuracy drop: 0.45 -> -0.03Improved
use adaptive hashing instead of uniform int8 hashing (ablation)RED++ · ImageNet MobileNet V2 (0.35)% accuracy drop: 3.96 -> 0.25Improved
use adaptive hashing instead of uniform on EfficientNet B1 (ablation)RED++ · ImageNet EfficientNet B1% accuracy drop: 14.32 -> 0.36Improved
use block alpha strategy instead of constant alpha (ablation)RED++ · Cifar10 ResNet 56pruning factor (%): 72.40 -> 77.68Improved

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