首页 > AI前沿 > Hierarchical Diffusion Transformer for End-to-End Image Generation

Hierarchical Diffusion Transformer for End-to-End Image Generation

arXiv机器学习 2026-11-17 23:47 8 阅读 查看原文

Pixel-space diffusion learns visual representations and image generation jointly in a single end-to-end training stage, without a separately pretrained visual encoder. However, plain pixel-space Transformers still trail leading two-stage latent diffusion models in generation quality and training efficiency. We hypothesize that part of this gap arises from retaining one token per patch throughout the backbone, which does not explicitly encourage a visual hierarchy that summarizes related content at progressively coarser levels.

To address this, we introduce Hierarchical Diffusion Transformers (HiT), which progressively