AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

MixMix: All You Need for Data-Free Compression Are Feature and Data Mixing

Yuhang Li, Feng Zhu, Ruihao Gong, Mingzhu Shen, Xin Dong, Fengwei Yu, Shaoqing Lu, Shi Gu

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Patient consistent distillation · 43 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 Data Mixing to Feature Mixing with m'=1 (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 64.46 -> 68.16Improved
add Data Mixing to Feature Mixing with m'=2 (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 68.87 -> 69.29Improved
add Data Mixing to Feature Mixing with m'=3 (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 69.49 -> 69.80Improved
add Data Mixing to Feature Mixing with m'=4 (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 69.54 -> 69.47No clear effect
mix more models in Feature Mixing (m' 1 -> 3) (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 64.46 -> 69.49Improved
mix more models in Feature Mixing (m' 3 -> 4) (ablation)MixMix · ResNet-50 structured pruningtop-1 accuracy: 69.49 -> 69.54No clear effect
use MixMix data instead of ZeroQ dataBRECQ PTQ · MobileNetV2 W4A4top-1 accuracy: 49.83 -> 64.01Improved
use MixMix data instead of DeepInversion datastructured pruning · MobileNetV2top-1 accuracy: 15.32 -> 42.47Improved
use MixMix data instead of DeepInversion datastructured pruning · MNasNettop-1 accuracy: 22.62 -> 43.41Improved
use MixMix data instead of KW dataLSQ QAT · MobileNetV2 W4A4top-1 accuracy: 66.07 -> 67.74Improved
add MixMix objective to DAFLDAFL · CIFAR10 ResNet-34 -> ResNet-18accuracy: 92.22 -> 94.79Improved
use MixMix data instead of DeepInversion datadata-free KD · CIFAR10 VGG11 -> VGG11accuracy: 90.78 -> 93.97Improved

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