面向快速自回归图像生成的并行 Jacobi 解码
Parallel Jacobi Decoding for Fast Autoregressive Image Generation
把 Jacobi 式投机解码的草稿序列从一维 token 链扩展到二维空间邻域，并配合调整注意力掩码抑制误差累积，可直接套用到 VAR、ImageGPT 等自回归图像生成器的推理阶段，在不重训的前提下换取数倍加速。
原文依据 we propose Parallel Jacobi Decoding (PJD), a training-free decoding approach that expands draft tokens in the two-dimensional spatial domain
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论自回归（AR）模型在生成高保真图像方面表现出卓越性能。然而，其固有的逐 token 顺序预测导致推理速度显著变慢。近期研究引入 Jacobi 式解码以加速自回归图像生成。扩展草稿序列最初能提升效率，但随着一维序列中的误差传播阻碍收敛，加速效果很快饱和。观察到图像具有强烈的局部空间相关性，我们提出并行 Jacobi 解码（PJD），一种免训练的解码方法，将草稿 token 扩展到二维空间域，从而实现高效的空间并行细化。PJD 调整注意力掩码以缓解误差累积并提升收敛稳定性。在多个数据集上的大量实验表明，PJD 在多种自回归图像生成模型上实现 4.8 倍至 6.4 倍加速，同时保持有竞争力的生成质量。
英文原文摘要 1024 字符
Autoregressive (AR) models have demonstrated remarkable performance in generating high-fidelity images. However, their inherently sequential next-token prediction leads to significantly slows inference. Recent studies have introduced Jacobi-style decoding to accelerate autoregressive image generation. Extending the draft sequence initially improves efficiency, yet the acceleration quickly saturates as error propagation in the one-dimensional sequence hinders convergence.Observing that images exhibit strong local spatial correlations, we propose Parallel Jacobi Decoding (PJD), a training-free decoding approach that expands draft tokens in the two-dimensional spatial domain to enable efficient spatially parallel refinement.PJD adjusts the attention mask to mitigate error accumulation and improve convergence stability. Extensive experiments on diverse datasets show that PJD achieves **4.8x-6.4x** acceleration across multiple autoregressive image generation models while maintaining competitive generation quality.