The authors trained MixTeacher, a dual-branch Diffusion Transformer that jointly models paired f8 and f32 latents extracted from the same image.
High-fidelity latent priors improve efficient high-resolution image generation
UltraDiff transfers guidance from lightly compressed image latents into a model operating in a substantially smaller latent space.
Big Tech
Jingjing Ren · Haitian Zheng · Haoyu Chen · Connelly Barnes · Zhifei Zhang · Yuqian Zhou · +2 more
The Hong Kong University of Science and Technology (Guangzhou) · Adobe Research
Research Digest··2 min read
Ren et al.
Why this paper
From Adobe Research and The Hong Kong University of Science and Technology (Guangzhou)
In one line
UltraDiff distills a joint f8-f32 diffusion teacher into an f32 student, approaching f8 image quality at f32 speed.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ✓Compute or model size stated (hardware A100)
- ·No stated limitations found
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§