AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Searching for MobileNetV3

Andrew Howard, Mark Sandler, Grace Chu, Liang-Chieh Chen, Bo Chen, Mingxing Tan, Weijun Wang, Yukun Zhu, Ruoming Pang, Vijay Vasudevan, Quoc V. Le, Hartwig Adam

ICCV 2019 · Architectures and efficiency › Neural architecture search · approach: Latency-aware architecture search · 10,855 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 h-swish instead of ReLU (ablation)MobileNetV3-Large · ImageNet valtop-1 accuracy: 74.5 -> 75.2Improved
use optimized h-swish implementation (ablation)MobileNetV3-Large · ImageNet val Pixel 1latency (ms): 57.5 -> 51.4Improved
enable h-swish from 16 channels layer instead of later (ablation)MobileNetV3-Large · ImageNet valtop-1 accuracy: 75.2 -> 75.4Mixed
enable h-swish only from 112 channels layer (ablation)MobileNetV3-Large · ImageNet valtop-1 accuracy: 75.2 -> 75.0Hurt
add LR-ASPP segmentation head instead of R-ASPP (ablation)MobileNetV3 + RF2 · Cityscapes valmIOU: 71.91 -> 72.37Improved
add LR-ASPP instead of R-ASPP (ablation)MobileNetV2 + RF2 · Cityscapes valmIOU: 72.56 -> 72.97Improved
reduce last-block channels by 2 (RF2) (ablation)MobileNetV3 R-ASPP · Cityscapes valmIOU: 72.64 -> 71.91Hurt
reduce segmentation head filters from 256 to 128 (ablation)MobileNetV3 LR-ASPP · Cityscapes valmIOU: 72.37 -> 72.36No clear effect
apply atrous convolution in last block (OS=16 vs 32) (ablation)MobileNetV3-Large · Cityscapes testmIOU: 72.0 -> 72.6Improved
reduce channels between C4 and C5 by 2 (ablation)MobileNetV3-Large SSDLite · COCO testmAP: 22.0 -> 22.0No clear effect
use MobileNetV3-Large instead of V2SSDLite · COCO testmAP: 22.1 -> 22.0No clear effect
use MobileNetV3-Large instead of MobileNetV2ImageNet classification · ImageNet valtop-1 accuracy: 72.0 -> 75.2Improved

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