Skip to content
NewA calmer, faster ZeroHour

Improved Distributional Diffusion Models

Hugging Face daily papersby Tommaso Martorella, Alexandre Galashov, Felix Krause ()1 min read1
AI summary · grok-4.7Summaries by Vani AI

Improved distributional diffusion models reach 2.38 FID on ImageNet-256 with one from-scratch DiT.

The paper makes Distributional Diffusion Models practical by deferring particle expansion to late transformer layers and using time-dependent scoring-rule schedules. A single from-scratch DiT-XL/2 latent model reaches 4.48 FID at 4 steps and 2.38 FID at 50 steps on class-conditional ImageNet-256, without a teacher, self-distillation, or Jacobian-vector products. FID does not degrade as the sampling budget grows from 4 to 50 evaluations, and the recipe also transfers to text-to-image generation. Code and pre-trained models are released by CompVis.

Key facts
  • Late-layer particle expansion reduces multi-particle training overhead.
  • Time-dependent scoring schedules replace one fixed hyperparameter trade-off.
  • DiT-XL/2 reaches 4.48 FID at 4 steps and 2.38 at 50.
  • The model is trained from scratch in one stage without a teacher.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at huggingface.co

Distributional Diffusion Models (DDMs) replace the standard mean-prediction denoiser with a distributional denoiser trained via a scoring rule objective, learning a stochastic approximation to p(x_1 mid x_t) rather than its conditional mean.

Hugging Face daily papers

By Tommaso Martorella, Alexandre Galashov, Felix Krause · huggingface.co

Read the full article at Hugging Face daily papers

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.