Metraipolis › archive › dc_09ef0c334e2e
VDOT++: Unified Few-Step Video Generation via Unbalanced Optimal Transport Distillation — arXiv:2610.03221
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.03221
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T09:08:02.823Z
- SHA-256 of the archived text
257fcb1beb7b7c7cc158f5b98d8b284ac5d523d31840ff22fef613b7cc3fc4ab- Stable URL of this page
https://metraipolis.com/documento/dc_09ef0c334e2e
How to cite
VDOT++: Unified Few-Step Video Generation via Unbalanced Optimal Transport Distillation — arXiv:2610.03221 — arXiv (en), https://arxiv.org/abs/2610.03221 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_09ef0c334e2e (Metraipolis archive dc_09ef0c334e2e, sha256 257fcb1beb7b7c7cc158f5b98d8b284ac5d523d31840ff22fef613b7cc3fc4ab)
BibTeX
@misc{dc09ef0c334e2e,
title = {VDOT++: Unified Few-Step Video Generation via Unbalanced Optimal Transport Distillation — arXiv:2610.03221},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.03221},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_09ef0c334e2e; sha256 257fcb1beb7b7c7cc158f5b98d8b284ac5d523d31840ff22fef613b7cc3fc4ab},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.03221. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Video creation spans text-to-video (T2V), image-to-video (I2V), and condition-based generation, yet video diffusion models remain costly because they repeatedly evaluate large backbones during sampling. Distribution matching distillation (DMD) reduces this cost, but its reverse Kullback--Leibler (KL) objective can provide unstable or incomplete guidance when the student and teacher distributions have limited overlap. VDOT addressed this issue by adding optimal transport distillation (OTD), whose explicit coupling supplies geometric directions for condition-based generation. Balanced OTD, however, performs full-mass matching between the spatial tokens of each corresponding student--teacher frame pair. This assumption weakens for T2V and I2V, where one condition admits many valid outputs and spatial content need not align across different realizations. We present VDOT++, a unified distillation framework that applies the same training recipe separately to generators for the three task families. It makes OTD robust to output diversity through an asymmetric unbalanced formulation that allows unreliable student tokens to carry less mass while maintaining coverage of the teacher tokens. An $\ell_1$ ground cost further replaces mean-based aggregation with a more mode-preserving weighted median that limits the influence of distant transport targets. The two changes respectively determine whom to match and how the selected targets should be aggregated. We additionally combine distribution matching and adversarial refinement through sequential backward passes, and exploit the decoupled score networks for cross-scale distillation, where larger score networks improve a compact generator. Experiments on UVCBench, VBench, VBench-I2V, and the VACE benchmark show that the resulting four-step generators are competitive with many-step teachers and strong few-step baselines across all three task families. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-10-05 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.03221 — 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 2026-10-06T08:21:44.221Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-10-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).