Improved Distributional Diffusion Models

Tommaso Martorella, Alexandre Galashov, Felix Krause, Stefan Andreas Baumann, Valentin De Bortoli, Arthur Gretton, Björn Ommer

Improved Distributional Diffusion Models: 5 upvotes on Hugging Face Daily Papers, #74 of 92 papers on 2026-09-30. Day-by-day upvote history.

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. However, scaling DDMs to modern image-generation settings faces two obstacles: (i) multi-particle training incurs overhead that scales with the number of particles, (ii) DDMs use globally fixed scoring rule hyperparameters, forcing a single trade-off across sampling budgets. We mitigate these limitations by deferring particle expansion to late transformer layers, and the hyperparameter trade-off by introducing time-dependent scoring rule schedules informed by the dynamical regimes of~Biroli2024. Combined with a DiT-based latent setup, these changes make DDM training practical on class-conditional ImageNet-256^2, achieving 4.48 FID at 4 steps and 2.38 at 50 steps with DiT-XL/2, from a single model trained from scratch in one stage, without a teacher, self-distillation or JVPs. The result is a stochastic few-step generator whose FID does not degrade as the sampling budget grows from 4 to 50 NFE, and the same recipe transfers to text-to-image generation. Code and pre-trained models available at https://github.com/CompVis/iDDM.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.