AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

NUWA-Infinity: Autoregressive over Autoregressive Generation for Infinite Visual Synthesis

Jian Liang, Chenfei Wu, Xiaowei Hu, Zhe Gan, Jianfeng Wang, Lijuan Wang, Zicheng Liu, Yuejian Fang, Nan Duan

NeurIPS 2022 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 25 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 autoregressive-over-autoregressive generation rather than the Taming or MaskGIT baselinesNUWA-Infinity · LHQC Text-to-Image HDBlock-FID: 24.33 -> 9.71; 38.89 -> 9.71Improved
Use autoregressive-over-autoregressive generation for images four times the training lengthNUWA-Infinity · LHQC Text-to-Image HD (4096×1024)Block-FID: 45.76 -> 15.65; 46.37 -> 15.65Improved
Use autoregressive-over-autoregressive generation instead of MaskGITNUWA-Infinity · LHQC Text-to-Image HDIS: 4.61 -> 4.98Improved
Use autoregressive-over-autoregressive generation instead of MaskGITNUWA-Infinity · LHQC Text-to-Image HDCLIP-SIM: 0.2754 -> 0.2807Improved
Use local autoregressive token generation instead of parallel non-autoregressive generation (ablation)NUWA-Infinity · LHQ-V Image Animation HDFVD: 368.16 -> 62.57Improved
Use local autoregressive token generation instead of progressive non-autoregressive generation (ablation)NUWA-Infinity · LHQ-V Image Animation HDFVD: 165.39 -> 62.57Improved
Add a text prompt to outpaintingNUWA-Infinity · LHQC Outpainting HD, down and up extendBlock-FID: 11.47 -> 9.84; 8.03 -> 7.43Improved
Add a text prompt to outpaintingNUWA-Infinity · LHQC Outpainting HD, right and left extendBlock-FID: 6.43 -> 6.45; 6.71 -> 6.72No clear effect
Feed positional embeddings before attention rather than after attention (ablation)NUWA-Infinity · LHQC Text-to-Image HDBlock-FID: 10.47 -> 10.05; 18.89 -> 17.78Improved
Add cached hidden states to the Nearby Context Pool (ablation)NUWA-Infinity · LHQC Text-to-Image HDBlock-FID: 15.80 -> 10.21; 38.32 -> 23.62Improved
Replace local autoregression with parallel non-autoregressive token generation (ablation)NUWA-Infinity · LHQC Text-to-Image HDBlock-FID: 10.05 -> 92.34; 17.78 -> 98.67Hurt
Replace local autoregression with progressive non-autoregressive token generation (ablation)NUWA-Infinity · LHQC Text-to-Image HDBlock-FID: 10.05 -> 19.86; 17.78 -> 38.59Hurt

Later papers that compared with this method

1 later accepted paper reports NUWA-Infinity as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.

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