Abstract
Posterior Matching Distillation trains a stochastic student to match a teacher diffusion model’s full posterior via tilted mean constraints, enabling high-quality few-step image generation without auxiliary score networks.
Diffusion distillation accelerates image and video generation by compressing a pretrained teacher's iterative sampling process into a few network evaluations. Distribution-based methods achieve this by matching the student's output distribution to the teacher's. This matching is often driven by reverse-KL objectives, whose mode-seeking bias favors dominant teacher modes at the expense of diversity across samples. We introduce Posterior Matching Distillation (PMD), a new member of the distribution-based distillation family that combines mode-covering posterior sampling with any-step generation. At zero population loss, PMD recovers the teacher's full posterior of clean images at each trained noise level. Its stochastic student supports any-step sampling budgets within the trained timestep grid without further training. To learn these posteriors, PMD matches exponentially reweighted averages of student samples conditioned on a fixed noisy state to corresponding teacher posterior means. Under Gaussian corruption, exponentially tilted posterior mean is a single teacher evaluation at a shifted input. We prove that, matching tilted means over an open neighborhood uniquely identifies the full posterior. PMD thus supervises the entire conditional distribution through a family of mean constraints, without any auxiliary student-score network. Experiments on SDXL and FLUX.2 [klein] 4B demonstrate that PMD achieves the best FID among 4-step methods on both models, while maintaining competitive or leading performance across prompt-alignment and perceptual-quality metrics.
Community
Hi authors, thanks for the work. A quick question: the Table 5 row for the default tilt radius U(0.25, 0.5) has the same diversity scores as PMD (FLUX.2) in Table 2, but its COCO-10k FID is 17.47, while Table 1 reports 24.12 for PMD on FLUX.2. Details in this GitHub issue: https://github.com/yesiltepe-hidir/PMD/issues/1
This is an automated message from the ResearchStudio team.
We created an interactive ResearchStudio Reel for this paper. It includes a visual poster, a video, and a blog, all available for download in editable formats.
Open the ResearchStudio Reel →
Download all files from Hugging Face
Please give this comment a thumbs up if you find the Reel helpful!
Want to explore or create Reels for more papers? Visit the ResearchStudio demo.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- VA-TDM: A Unified View for Improving Few-Step Diffusion Models by Velocity Adversarial Distillation (2026)
- Intrinsic Distribution Matching (2026)
- FisherDistill: Diffusion Distribution Distillation via Closed-Form Fisher Critics (2026)
- DyMD: Preserving Interaction Dynamics through Distribution Matching Distillation in Few-Step Video World Models (2026)
- eDMD: Endpoint Distribution Matching for Few-Step Diffusion Distillation (2026)
- A First-Step Bias Restores Lost Diversity in Diffusion Model Distillation (2026)
- Direct Forcing: Single-Stage Training with Representation-Space Supervision for Few-Step Autoregressive Video Generation (2026)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2611.26629 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
