STARFlow:扩展潜在归一化流以实现高分辨率图像合成
We present STARFlow, a scalable generative model based on normalizing flows that achieves strong performance in high-resolution image synthesis. The core of STARFlow is Transformer Autoregressive...
STARFlow是一种基于归一化流的生成模型,专注于高分辨率图像合成。其核心是Transformer自回归流(TARFlow),结合了归一化流的表达能力与自回归Transformer的结构建模能力。通过深浅设计、在预训练自编码器的潜在空间建模及新颖的引导算法,STARFlow显著提升了可扩展性和样本质量,接近最先进的扩散模型表现。
