AI Paper Atlas

Research map / Image generation / GANs

Online Multi-Granularity Distillation for GAN Compression

Yuxi Ren, Jie Wu, Xuefeng Xiao, Jianchao Yang

ICCV 2021 · Image generation › GANs · approach: GAN latent optimization and evaluation · 50 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
use online distillation instead of offline two-stage (ablation)OMGD (U-Net) · CityscapesmIoU: 45.76 -> 48.91Improved
use online distillation instead of offline two-stage (ablation)OMGD (U-Net) · edges->shoesFID: 26.19 -> 25.00Improved
use online distillation instead of offline two-stage (ablation)OMGD (Res-Net) · horse->zebraFID: 77.09 -> 51.92Improved
add deeper teacher to the wider teacher (ablation)OMGD (U-Net) · edges->shoesFID: 33.88 -> 25.00Improved
add deeper teacher to the wider teacher (ablation)OMGD (U-Net) · CityscapesmIoU: 44.04 -> 48.91Improved
add channel distillation of intermediate layers (ablation)OMGD (U-Net) · edges->shoesFID: 26.62 -> 25.00Improved
add channel distillation of intermediate layers (ablation)OMGD (U-Net) · CityscapesmIoU: 48.12 -> 48.91Improved
add channel distillation of intermediate layers (ablation)OMGD (Res-Net) · horse->zebraFID: 61.21 -> 51.92Improved
add channel distillation of intermediate layers (ablation)OMGD (Res-Net) · summer->winterFID: 75.47 -> 73.79Improved
train student with discriminator and GAN loss instead of discriminator-free (ablation)OMGD (U-Net) · edges->shoesFID: 25.00 -> 73.44Hurt
train student with discriminator and GAN loss instead of discriminator-free (ablation)OMGD (U-Net) · CityscapesmIoU: 48.91 -> 31.79Hurt
add learn-best teacher selection strategy (ablation)OMGD (Res-Net) · horse->zebraFID: 63.15 -> 52.00Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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