AI Paper Atlas

Research map / Image generation / Super-resolution

VARestorer: One-Step VAR Distillation for Real-World Image Super-Resolution

Yixuan Zhu, Shilin Ma, Haolin Wang, Ao Li, Yanzhe Jing, Yansong Tang, Lei Chen, Jiwen Lu, Jie Zhou

ICLR 2026 · Image generation › Super-resolution · approach: Efficient lightweight super-resolution · 1 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 one-step VAR distillation instead of multi-step ControlNet VAR (ablation)VARestorer · DIV2K-ValMUSIQ: 62.22 -> 72.32Improved
add one-step VAR distillation instead of multi-step ControlNet VAR (ablation)VARestorer · DIV2K-ValCLIPIQA: 0.4794 -> 0.7669Improved
add cross-scale full attention instead of causal attention (ablation)VARestorer · DIV2K-ValMUSIQ: 63.72 -> 72.32Improved
add cross-scale full attention instead of causal attention (ablation)VARestorer · DIV2K-ValCLIPIQA: 0.3910 -> 0.7669Improved
add cross-scale full attention instead of causal attention (ablation)VARestorer · DIV2K-ValLPIPS: 0.4224 -> 0.3131Improved
add KL distribution matching loss (ablation)VARestorer · DIV2K-ValMUSIQ: 69.73 -> 72.32Improved
add KL distribution matching loss (ablation)VARestorer · DIV2K-ValCLIPIQA: 0.6682 -> 0.7669Improved
add KL distribution matching loss (ablation)VARestorer · DIV2K-ValNIQE: 4.372 -> 4.410Hurt
add BLIP text prompt at inference (ablation)VARestorer · DIV2K-ValMUSIQ: 67.05 -> 72.32Improved
use LoRA tuning instead of full fine-tuning (ablation)VARestorer · DIV2K-ValMUSIQ: 70.58 -> 72.32Improved
train at 1024 resolution instead of 512 (ablation)VARestorer · DIV2K-ValMUSIQ: 72.32 -> 72.11No clear effect
train longer (20K and 25K steps vs 10K) (ablation)VARestorer · DIV2K-Valall metricsNo clear effect

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