Metraipolis › archive › dc_15cbaba96119

StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models — arXiv:2611.04386

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04386
License (as archived)
CC0-1.0
Archived
2026-11-04T12:56:21.761Z
SHA-256 of the archived text
5d593a470d40a3e2fb32d9ec5a62d97d39c3338dd16c52afe9ee71df3c721cc0
Stable URL of this page
https://metraipolis.com/documento/dc_15cbaba96119

How to cite

StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models — arXiv:2611.04386 — arXiv (en), https://arxiv.org/abs/2611.04386 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_15cbaba96119 (Metraipolis archive dc_15cbaba96119, sha256 5d593a470d40a3e2fb32d9ec5a62d97d39c3338dd16c52afe9ee71df3c721cc0)
BibTeX
@misc{dc15cbaba96119,
  title = {StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models — arXiv:2611.04386},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04386},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_15cbaba96119; sha256 5d593a470d40a3e2fb32d9ec5a62d97d39c3338dd16c52afe9ee71df3c721cc0},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04386. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A deployable text-to-image diffusion model must hold several capabilities at once, such as instruction following, text rendering and human preference, and must sample in a few steps. Capabilities and few-step sampling are built in separate stages: a full-step model first consolidates capabilities through multi-teacher on-policy distillation, then serves as a fixed teacher for step distillation. Every full-step model we evaluate scores lower once compressed, and the second stage cannot begin until the first finishes. We introduce StepX-OPD, which merges the two stages: one student acquires every capability while learning to sample in a few steps, so frozen specialists supervise it on the states its own rollout visits rather than on a longer trajectory it never runs. Because a decomposition of the rollout's KL weights velocity mismatch most heavily near the clean end, we also regress the student's velocity on the low-noise part of the rollout. Separately, we find that capability integration makes models write prompt words into images on prompts that never asked for text, and we give a variant that lowers it. We evaluate on SD3.5-Medium and FLUX.1-dev across instruction following, text rendering, image quality and human preference. On both architectures, StepX-OPD attains the highest normalized average score among the four-step methods compared and exceeds the forty-step Flow-OPD model, and on SD3.5-Medium it trains with about 37% fewer GPU-hours than sequential DiffusionOPD followed by step distillation.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2611.04386 — 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 2026-11-04T12:33:47.135Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).