Distilling Specialized Orders for Visual Generation

[2504.17069] Distilling Specialized Orders for Visual Generation Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Computer Vision and Pattern Recognition arXiv:2504.17069 (cs) [Submitted on 23 Apr 2025 ( v1 ), last revised 8 Apr 2026 (this version, v2)] Title: Distilling Specialized Orders for Visual Generation Authors: Rishav Pramanik , Amin Sghaier , Masih Aminbeidokhti , Juan A. Rodriguez , Antoine Poupon , David Vazquez , Christopher Pal , Zhaozheng Yin , Marco Pedersoli View a PDF of the paper titled Distilling Specialized Orders for Visual Generation, by Rishav Pramanik and 8 other authors View PDF HTML (experimental) Abstract: Autoregressive (AR) image generators are becoming increasingly popular due to their ability to produce high-quality images and their scalability. Typical AR models are locked onto a specific generation order, often a raster-scan from top-left to bottom-right; this prohibits multi-task flexibility (inpainting, editing, outpainting) without retraining. Any-order AR models address this by learning to generate under arbitrary patch orderings, but at the cost of increased complexity and lower performance. In this paper, we present Ordered Autoregressive (OAR) generation, a self-distillation pipeline that first trains an any-order AR model, then extracts specialized generation orders from the model's own confidence scores, and fine-tunes on these orders. This achieves two goals: 1) improved generation quality by redirecting capacity from learning all $N!$ orderings to a single specialized path, and 2) preserved flexibility of any-order models. On ImageNet $256\times 256$, OAR improves FID from 2.39 to 2.17 over the any-order baseline, with consistent gains on Fashion Products and CelebA-HQ. OAR supports zero-shot inpainting and outpainting without retraining, and human evaluation shows 64% preference over the baseline. The pipeline requires only lightweight fine-tuning on a pretrained any-order model, with no architectural changes or additional annotations. Subjects: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI) Cite as: arXiv:2504.17069 [cs.CV] (or arXiv:2504.17069v2 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2504.17069 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Rishav Pramanik [ view email ] [v1] Wed, 23 Apr 2025 19:33:58 UTC (12,286 KB) [v2] Wed, 8 Apr 2026 18:21:06 UTC (12,995 KB) Full-text links: Access Paper: View a PDF of the paper titled Distilling Specialized Orders for Visual Generation, by Rishav Pramanik and 8 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.CV < prev | next > new | recent | 2025-04 Change to browse by: cs cs.AI References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Qwen-Image-Flash: Rethinking the Training Recipe for Few-Step Distillation#1084232
• Mira-Scene: Pixel-Aligned Layouts for Generative 3D Scene#1054722
• UniSonate: A Unified Model for Speech, Music, and Sound Effect Generation with Text Instructions#134610
• IB-Flow: Information Bottleneck-Guided CFG Distillation for Few-Step Text-to-Image Generation#1063964
• JoLT: Joint Latent Trajectories for Context-Guided High-Resolution Tiled Generation#615362