AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Next Patch Prediction for AutoRegressive Visual Generation

Yatian Pang, Peng Jin, Shuo Yang, Bin Zhu, Bin Lin, Chaoran Feng, Zhenyu Tang, Liuhan Chen, Francis E. H. Tay, Ser-Nam Lim, Harry Yang, Li Yuan

AAAI 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 30 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
train with next patch prediction (2x2 patches then 1x1) instead of next token predictionLlamaGen-B · ImageNet 256x256FID: 5.46 -> 4.47Improved
train with next patch predictionLlamaGen-L · ImageNet 256x256FID: 3.80 -> 2.76Improved
train with next patch predictionLlamaGen-XL · ImageNet 256x256FID: 3.39 -> 2.65Improved
train with next patch predictionLlamaGen-XXL · ImageNet 256x256FID: 3.10 -> 2.54Improved
train with next patch prediction (training compute)LlamaGen-L · ImageNet 256x256training GFLOPs: 83.54 -> 47.95Improved
enlarge first-stage patch size from 2x2 to 4x4 (ablation)NPP-B · ImageNet 256x256FID: 4.47 -> 4.92Hurt
enlarge first-stage patch size from 2x2 to 4x4 (ablation)NPP-L · ImageNet 256x256FID: 2.76 -> 2.89Hurt
raise segment factor lambda from 1/2 to 4/5 (ablation)NPP-L · ImageNet 256x256FID: 2.76 -> 2.92Hurt
raise segment factor lambda from 1/2 to 4/5 (cost) (ablation)NPP-L · ImageNet 256x256training cost: 0.625x -> 0.4xImproved
use more patch levels (#L=2 to 4) (ablation)NPP-L · ImageNet 256x256FID: 2.76 -> 2.84Hurt
use more patch levels (#L=2 to 4) (ablation)NPP-B · ImageNet 256x256FID: 4.47 -> 4.68Hurt
use NPP-XL instead of LlamaGen-XL-384NPP-XL · ImageNet 256x256FID: 2.62 vs 2.65No clear effect

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