Metraipolis › archive › dc_2ca57a4630d5

Consistent Distribution Matching for Data-Free Diffusion Distillation — arXiv:2610.09221

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09221
License (as archived)
CC0-1.0
Archived
2026-10-08T11:10:43.800Z
SHA-256 of the archived text
fb8f4f6cc67722735b6f8b0c781c7c1002697502d66959d94d848590b5682c59
Stable URL of this page
https://metraipolis.com/documento/dc_2ca57a4630d5

How to cite

Consistent Distribution Matching for Data-Free Diffusion Distillation — arXiv:2610.09221 — arXiv (en), https://arxiv.org/abs/2610.09221 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2ca57a4630d5 (Metraipolis archive dc_2ca57a4630d5, sha256 fb8f4f6cc67722735b6f8b0c781c7c1002697502d66959d94d848590b5682c59)
BibTeX
@misc{dc2ca57a4630d5,
  title = {Consistent Distribution Matching for Data-Free Diffusion Distillation — arXiv:2610.09221},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09221},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2ca57a4630d5; sha256 fb8f4f6cc67722735b6f8b0c781c7c1002697502d66959d94d848590b5682c59},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09221. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Flow and diffusion models suffer from slow inference due to computationally expensive numerical integration. Distillation provides a promising way for a student model to learn from a teacher's dynamics, enabling one-step or few-step generation. However, existing methods often depend on curated distillation datasets, costly teacher rollouts, or auxiliary proxy networks, which complicate model training and scaling. In this work, we propose Consistent Distribution Matching, a simulation-free and data-free distillation method for accelerating diffusion and flow models while preserving strong generative capacity. Our key insight is to unify sample generation and score estimation with one student network. Thus, our framework uses only two models, a frozen teacher and a trainable student, and optimizes one objective. We prove that minimizing our objective indicates Wasserstein convergence of the student flow-map pushforwards to the teacher marginals. On ImageNet 256$\times$256, our method attains an FID of 2.04 with a single function evaluation (1-NFE) and a 4-NFE FID of 1.37 within 40 epochs of training, surpassing the state-of-the-art distillation baselines without data. Our code code and model are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.09221 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T10:14:06.062Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).