研究人员引入了两种新方法来增强自回归视觉生成模型。第一种称为 Prologue，通过预置一组仅为生成而训练的小型序言令牌来解决重建-生成差距，从而显著提高了 ImageNet 上的图像质量。第二种，视觉隐式自回归建模 (VIAR)，嵌入了一个隐式平衡层，以减少计算内存并允许在推理过程中进行计算控制，以更少的参数和更高的效率实现了有竞争力的结果。 AI
影响 这些论文引入的新技术可能导致更高效、更高质量的图像生成模型。
排序理由 两篇新的学术论文提出了改进自回归视觉生成的新颖方法。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.06137v1 Announce Type: cross Abstract: In this work, we propose Prologue, an approach to bridging the reconstruction-generation gap in autoregressive (AR) image generation. Instead of modifying visual tokens to satisfy both reconstruction and generation, Prologue gener…
In this work, we propose Prologue, an approach to bridging the reconstruction-generation gap in autoregressive (AR) image generation. Instead of modifying visual tokens to satisfy both reconstruction and generation, Prologue generates a small set of prologue tokens prepended to t…
arXiv:2605.01220v1 Announce Type: new Abstract: Visual Autoregressive Modeling (VAR) based on next-scale prediction achieves strong generation quality, but their explicit deep stacks fix the amount of computation per scale and inflate memory at high resolutions. We introduce Visu…