AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Towards Accurate Network Quantization with Equivalent Smooth Regularizer

Kirill Solodskikh, Vladimir Chikin, Ruslan Aydarkhanov, Dehua Song, Irina Zhelavskaya, Jiansheng Wei

ECCV 2022 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization · 3 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 multistep lambda_w scheduling (1,10,100) instead of constant lambda_w (ablation)QSin ResNet-20 · CIFAR-10 (4-bit)accuracy (%): 91.1 -> 91.7Improved
add multistep lambda_w scheduling instead of constant lambda_w=100 (ablation)QSin ResNet-20 · CIFAR-10 (8-bit)accuracy (%): 91.8 -> 91.9Improved
add weight regularizer lambda_w=1 (with lambda_a=0) (ablation)QSin ResNet-20 · CIFAR-10 (4-bit)accuracy (%): 88.9 -> 90.0Improved
add activation regularizer lambda_a=1 (lambda_w=0) (ablation)QSin ResNet-20 · CIFAR-10 (4-bit)accuracy (%): 88.9 -> 89.2Improved
raise lambda_w from 1 to 100 (ablation)QSin ResNet-20 · CIFAR-10 (4-bit)accuracy (%): 90.2 -> 91.1Improved
regularizers on 8-bit (any lambda settings) (ablation)QSin ResNet-20 · CIFAR-10 (8-bit)accuracy (%): 91.6 -> 91.9No clear effect
replace MSQE regularizer with smooth QSin regularizerResNet-18 quantization · ImageNet (4-bit)top-1 accuracy (%): 67.3 -> 69.7Improved
replace MSQE regularizer with QSinMobileNet-V2 quantization · ImageNet (4-bit)top-1 accuracy (%): 67.4 -> 68.7Improved
replace SinReQ sinusoidal regularizer with QSinResNet-18 quantization · ImageNet (4-bit)top-1 accuracy (%): 64.63 -> 69.7Improved
use QSin instead of LSQ (round-free, no STE)MobileNet-V2 quantization · ImageNet (4-bit)top-1 accuracy (%): 68.1 -> 68.7Improved
use QSin instead of LSQResNet-18 quantization · ImageNet (4-bit)top-1 accuracy (%): 69.8 -> 69.7No clear effect
use QSin instead of MSQE for 8-bitEDSR 4x SR · Urban100 (8-bit)PSNR (dB): 25.9 -> 26Improved

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