AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

LazyMAR: Accelerating Masked Autoregressive Models via Feature Caching

Feihong Yan, Qingyan Wei, Jiayi Tang, Jiajun Li, Yulin Wang, Xuming Hu, Huiqi Li, Linfeng Zhang

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Accelerating autoregressive image generation · 10 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 token cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalspeedup: 1.00 -> 1.59Improved
add token cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalFID: 1.59 -> 1.62No clear effect
add condition cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalspeedup: 1.00 -> 1.87Improved
add condition cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalFID: 1.59 -> 1.67No clear effect
add condition cache on top of token cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalspeedup: 1.59 -> 2.83Improved
add token cache on top of condition cache (ablation)MAR-H, 64 steps · ImageNet 256x256 class-conditionalspeedup: 1.87 -> 2.83Improved
select tokens by max similarity at 3rd layer instead of random caching (ablation)LazyMAR, MAR-H 64 steps · ImageNet 256x256 class-conditionalFID: 1.95 -> 1.69Improved
select tokens by max similarity instead of min similarity (ablation)LazyMAR, MAR-H 64 steps · ImageNet 256x256 class-conditionalFID: 2.10 -> 1.69Improved
cache decoded tokens instead of unpredicted tokens (ablation)MAR-H 64 steps with condition cache · ImageNet 256x256 class-conditionalFID: 3.61 -> 1.79Improved
use similarity-based selection instead of caching only predicted tokens (ablation)MAR-H 64 steps with condition cache · ImageNet 256x256 class-conditionalFID: 1.79 -> 1.69Improved
use LazyMAR instead of reducing decoding steps from 64 to 32 (similar speed) (ablation)MAR-H · ImageNet 256x256 class-conditionalFID: 1.92 -> 1.69Improved
use LazyMAR instead of reducing decoding steps to 16MAR-B · ImageNet 256x256 class-conditionalFID: 4.10 -> 2.64Improved

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