AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

ARVAR: Accelerating Visual Autoregressive Model via Attention Retrospect

Jiedong Zhuang, Lu Lu, Ming Dai, Jian Chen, Qiang Liu, Haoji Hu

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 0 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
replace frequency-based token selection with attention-saliency (self-attn) selection (ablation)ARVAR on HART · MJHQ30KFID: 28.19 -> 27.67Improved
use random token pruning instead of saliency (ablation)ARVAR on HART · MJHQ30KFID: 27.67 -> 31.98Hurt
use self-attention part rather than full attention for saliency (ablation)ARVAR on HART · MJHQ30KFID: 28.23 -> 27.67Improved
add Attention Retrospect alone (ablation)HART · MJHQ30KFID / speedup: 30.75 -> 27.56, 1.46xImproved
add Branch Halving alone (ablation)HART · MJHQ30KFID / speedup: 30.75 -> 28.01, 1.43xImproved
combine AR and BH (ablation)HART · MJHQ30Kspeedup: 1.46x -> 1.56x speedup, FID 27.56 -> 27.67, SSIM 0.7232 -> 0.6904Mixed
ARVAR instead of FastVARHART · MJHQ30KSSIM / FID: SSIM 0.6689 -> 0.6904, FID 28.19 -> 27.67Improved
ARVAR instead of FastVARInfinity-2B · MJHQ30KPSNR: 20.47 -> 22.73Improved
apply ARVAR to HART on HPSv2.1HART · HPSv2.1HPSv2.1: 28.72 -> 27.67Hurt
apply ARVAR to Infinity-2BInfinity-2B · GenEvaloverall: 0.68 -> 0.69No clear effect
prune at later step with higher ratio (75% at S14 vs 50% at S13) (ablation)ARVAR on HART · MJHQ30KFID / speedup: 28.12 -> 27.67 FIDImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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