AI Paper Atlas

Research map / Image generation / Super-resolution

Universally Slimmable Networks and Improved Training Techniques

Jiahui Yu, Thomas S. Huang

ICCV 2019 · Image generation › Super-resolution · approach: CNN and transformer SR architectures · 390 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 both smallest and largest width to the sampled widths (sandwich rule) instead of 3 random widths (ablation)US-MobileNet v1 · ImageNet val (AVG over widths)top-1 error (%): 38.20 -> 36.45Improved
always train the smallest width in addition to 2 random widths (ablation)US-MobileNet v1 · ImageNet val (AVG over widths)top-1 error (%): 38.20 -> 36.73Improved
always train the largest width in addition to 2 random widths (ablation)US-MobileNet v1 · ImageNet val (AVG over widths)top-1 error (%): 38.20 -> 38.70Mixed
add inplace distillation from the largest width (ablation)US-MobileNet v1 · ImageNet val (FLOPs-accuracy spectrum)top-1 error (%)Improved
add inplace distillation from the largest width (ablation)US-WDSR · DIV2K x2 super-resolutionPSNRImproved
replace individually trained networks with a single US-NetMobileNet v1 · ImageNet val (AVG over widths)top-1 error (%): 36.9 -> 34.7Improved
replace 4-switch slimmable network with a US-NetMobileNet v2 · ImageNet val (AVG over widths)top-1 error (%): 34.1 -> 32.9Improved
replace individually trained models with a US-WDSRWDSR · DIV2K x2 super-resolutionPSNR: 0.01 lower PSNRHurt
replace individually trained networks with a US-NetAtari Breakout PPO · BreakoutNoFrameskip-v4mean episode rewardImproved
use exact average instead of moving average for BN post-statistics (ablation)US-MobileNet v1 · ImageNet val (0.25x)top-1 error (%): 44.4 -> 44.3No clear effect
compute BN post-statistics on 1k samples instead of 1.28M (ablation)US-MobileNet v1 · ImageNet val (0.25x)top-1 error (%): 44.3 -> 44.4No clear effect
add output averaging by number of input channels (ablation)US-MobileNet v1 · ImageNet val (AVG over widths)top-1 error (%): 34.7 -> 34.3Improved

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