Research map / Image generation / Autoregressive and latent generation
InfinityGAN: Towards Infinite-Pixel Image Synthesis
Chieh Hubert Lin, Hsin-Ying Lee, Yen-Chi Cheng, Sergey Tulyakov, Ming-Hsuan Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use InfinityGAN as the image prior in the In&Out outpainting pipeline | In&Out · Place365 | FID: 23.57 -> 9.11 | Improved |
| Use InfinityGAN as the image prior in the In&Out outpainting pipeline | In&Out · Place365 | IS: 7.18 -> 6.78 | Hurt |
| Use InfinityGAN as the image prior in the In&Out outpainting pipeline | In&Out · Flickr-Scenery | FID: 30.34 -> 15.31 | Improved |
| Use InfinityGAN as the image prior in the In&Out outpainting pipeline | In&Out · Flickr-Scenery | IS: 7.16 -> 7.19 | Improved |
| Increase parallel patch batch size from 1 to 128 using 8 GPUs | InfinityGAN · 8192×8192 image inference | speed up: × 1.00 -> × 7.20 | Improved |
| Remove zero-padding without adding a replacement source of positional information (ablation) | StyleGAN2+NCI · Flickr-Landscape | FID: 4.19 -> 86.76 | Hurt |
| Remove zero-padding without adding a replacement source of positional information (ablation) | StyleGAN2+NCI · Flickr-Landscape | 4× ScaleInv FID: 79.83 -> 211.22 | Hurt |
| Remove zero-padding without adding a replacement source of positional information (ablation) | StyleGAN2+NCI · Flickr-Landscape | 8× ScaleInv FID: 189.65 -> 272.80 | Hurt |
| Add the structure synthesizer to provide positional information (ablation) | StyleGAN2+NCI+PFG · Flickr-Landscape | FID: 86.76 -> 11.03 | Improved |
| Add the structure synthesizer to provide positional information (ablation) | StyleGAN2+NCI+PFG · Flickr-Landscape | 4× ScaleInv FID: 211.22 -> 61.41 | Improved |
| Add the structure synthesizer to provide positional information (ablation) | StyleGAN2+NCI+PFG · Flickr-Landscape | 8× ScaleInv FID: 272.80 -> 121.18 | Improved |
Later papers that compared with this method
3 later accepted papers report InfinityGAN as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Patched Denoising Diffusion Models For High-Resolution Image Synthesis (ICLR 2024 · better on 1 of 2 shared results)
- MaskGIT: Masked Generative Image Transformer (CVPR 2022 · better on 1 of 1 shared result)
- Vector Quantized Image-to-Image Translation (ECCV 2022 · better on 1 of 1 shared result)
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.
- In&Out — InOut: Diverse Image Outpainting via GAN Inversion (CVPR 2022) · better on 3 of 4 shared results
- SinGAN — SinGAN: Learning a Generative Model From a Single Natural Image (ICCV 2019) · better on 2 of 3 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
- Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis (CVPR 2025 · similarity 0.88)
- NUWA-Infinity: Autoregressive over Autoregressive Generation for Infinite Visual Synthesis (NeurIPS 2022 · similarity 0.87)
- Rethinking Generative Image Pretraining: How Far Are We From Scaling Up Next-Pixel Prediction? (ICML 2026 · similarity 0.87)
- Any-Resolution Training for High-Resolution Image Synthesis (ECCV 2022 · similarity 0.87)
- Novel View Synthesis with Pixel-Space Diffusion Models (CVPR 2025 · similarity 0.86)
- Scaling Up GANs for Text-to-Image Synthesis (CVPR 2023 · similarity 0.86)
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.
