AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Vertical Layering of Quantized Neural Networks for Heterogeneous Inference

Wu Hai, Ruifei He, Haoru Tan, Xiaojuan Qi, Kaibin Huang

TPAMI 2023 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization

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 individual BN layers per bit-width instead of shared BN (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 2/2)Top-1 accuracy (%): 33.5 -> 66.7Improved
use individual BN layers instead of individual statistics only (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 3/3)Top-1 accuracy (%): 68.8 -> 69.9Improved
add positive compensation in downsampling (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 2/2)Top-1 accuracy (%): 61.3 -> 66.7Improved
add positive compensation in downsampling (ablation)vertical-layered model · ImageNet MobileNetV2 (W/A 2/2)Top-1 accuracy (%): 40.4 -> 50.6Improved
use unitary scalarization instead of multi-gradient descent (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 4/4)Top-1 accuracy (%): 69.8 -> 70.6Improved
add cosine-similarity self-KD (ablation)vertical-layered model · ImageNet ResNet-50 (W/A 2/2)Top-1 accuracy (%): 72.4 -> 73.2Improved
add KL-divergence self-KD instead of cosine similarity (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 4/4)Top-1 accuracy (%): 70.6 -> 70.4No clear effect
replace full-precision first and last layers with bit-width specific ones (ablation)vertical-layered model · ImageNet ResNet-50 (W/A 2/2)Top-1 accuracy (%): 72.4 -> 71.0Hurt
replace full-precision first and last layers with bit-width specific ones (ablation)vertical-layered model · ImageNet ResNet-18 (W/A 2/2)Top-1 accuracy (%): 66.7 -> 66.9No clear effect
use joint QAT instead of ascending layer trainingvertical-layered model · ImageNet ResNet-18 (W/A 3/3)Top-1 accuracy (%): 67.9 -> 69.9Improved
use joint training instead of QAT + fine-tuningscalable bit-width model · ImageNet ResNet-50 (W/A 2/2)Top-1 accuracy (%): 60.1 -> 72.4Improved
use vertical-layered joint model instead of bit-width specific LSQquantized ResNet-18 · CIFAR-100 (W/A 2/2)Top-1 accuracy (%): 76.9 -> 76.3Hurt

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