BriefGPT.xyz
Jun, 2024
无向量量化的自回归图像生成
Autoregressive Image Generation without Vector Quantization
HTML
PDF
Tianhong Li, Yonglong Tian, He Li, Mingyang Deng, Kaiming He
TL;DR
我们提出了使用扩散过程模型每个令牌的概率分布,从而可以将自回归模型应用于连续值空间的方法,并通过定义扩散损失函数来替代离散化的令牌化。通过消除向量量化,我们的图像生成器在享受序列建模的速度优势的同时取得了强大的结果,并希望该方法能促进在其他连续值领域和应用中使用自回归生成。
Abstract
Conventional wisdom holds that
autoregressive models
for
image generation
are typically accompanied by vector-quantized tokens. We observe that while a discrete-valued space can facilitate representing a categori
→