paper-with-me

Papers

ZipAR: Accelerating Auto-regressive Image Generation through Spatial Locality

2024-12-05 · Yefei He, Feng Chen, Yuanyu He, Shaoxuan He, Hong Zhou, Kaipeng Zhang, Bohan Zhuang

In this paper, we propose ZipAR, a training-free, plug-and-play parallel decoding framework for accelerating auto-regressive (AR) visual generation. The motivation stems from the observation that images exhibit local structures, and spatially distant regions tend to have minimal interdependence. Given a partially decoded set of visual tokens, in addition to the original next-token prediction scheme in the row dimension, the tokens corresponding to spatially adjacent regions in the column dimension can be decoded in parallel, enabling the ``next-set prediction'' paradigm. By decoding multiple tokens simultaneously in a single forward pass, the number of forward passes required to generate an image is significantly reduced, resulting in a substantial improvement in generation efficiency. Experiments demonstrate that ZipAR can reduce the number of model forward passes by up to 91% on the Emu3-Gen model without requiring any additional retraining. Code is available here: https://github.com/ThisisBillhe/ZipAR.

📄 PDF Abstract BibTeX arXiv:2412.04062

Code (1)

ThisisBillhe/ZipAR 공식 구현 pytorch

Tasks

Image Generation

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding

2024-10-02 · Yao Teng, Han Shi, Xian Liu, Xuefei Ning 외

The current large auto-regressive models can generate high-quality, high-resolution images, but these models require hundreds or even thousands of steps of next-token prediction during inference, resulting in substantial…

Image GenerationText to Image GenerationText-to-Image Generation

Lformer: Text-to-Image Generation with L-shape Block Parallel Decoding

2023-03-07 · Jiacheng Li, Longhui Wei, Zongyuan Zhan, Xin He 외

Generative transformers have shown their superiority in synthesizing high-fidelity and high-resolution images, such as good diversity and training stability. However, they suffer from the problem of slow generation since…

DiversityImage GenerationText to Image GenerationText-to-Image Generation

Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generation

2025-10-10 · Yao Teng, Fuyun Wang, Xian Liu, Zhekai Chen 외 arxiv

As a new paradigm of visual content generation, autoregressive text-to-image models suffer from slow inference due to their sequential token-by-token decoding process, often requiring thousands of model forward passes to…

Text-to-Image Generation

PathRelax: Parallel-Path Relaxed Speculative Jacobi Decoding for Accelerating Auto-Regressive Text-to-Image Generation

2026-06-09 · Haodong Lei, Hongsong Wang, Bingxuan Dai, Pan Zhou arxiv

The growing need for high-resolution image generation in autoregressive text-to-image models has resulted in extended token sequences, significantly increasing computational costs and inference times. However, existing s…

Text-to-Image Generation

LSRS: Latent Scale Rejection Sampling for Visual Autoregressive Modeling

2025-12-03 · Hong-Kai Zheng, Piji Li arxiv

Visual Autoregressive (VAR) modeling approach for image generation proposes autoregressive processing across hierarchical scales, decoding multiple tokens per scale in parallel. This method achieves high-quality generati…

Computational EfficiencyImage Generation