Metraipolis › archive › dc_7d3c72f26fea

CLiS: Plug-and-Play Attention Acceleration for Video Diffusion on TPUs via Certified Stabilization — arXiv:2611.00444

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.00444
License (as archived)
CC0-1.0
Archived
2026-11-03T18:35:12.798Z
SHA-256 of the archived text
118e9e1c601e567fa60ed87274d4018c22e99399c6925752b2cc9a6a76f96374
Stable URL of this page
https://metraipolis.com/documento/dc_7d3c72f26fea

How to cite

CLiS: Plug-and-Play Attention Acceleration for Video Diffusion on TPUs via Certified Stabilization — arXiv:2611.00444 — arXiv (en), https://arxiv.org/abs/2611.00444 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7d3c72f26fea (Metraipolis archive dc_7d3c72f26fea, sha256 118e9e1c601e567fa60ed87274d4018c22e99399c6925752b2cc9a6a76f96374)
BibTeX
@misc{dc7d3c72f26fea,
  title = {CLiS: Plug-and-Play Attention Acceleration for Video Diffusion on TPUs via Certified Stabilization — arXiv:2611.00444},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.00444},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7d3c72f26fea; sha256 118e9e1c601e567fa60ed87274d4018c22e99399c6925752b2cc9a6a76f96374},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00444. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

High-resolution video diffusion relies on attention over long token sequences, making attention efficiency critical to generation latency. Our analysis identifies repeated maximum reductions and accumulator rescaling in online softmax as a bottleneck in long-context TPU attention. We propose CLiS (Certified Loop-invariant Stabilization), a training-free, plug-and-play optimization for video diffusion inference. CLiS fixes per-row stabilizers before the KV loop and certifies numerical safety per head, falling back to online softmax when certification fails. It integrates into existing fused-attention kernels without changing model weights or architecture. On Wan 2.2 and LTX-2 across TPU v6e and TPU7x, CLiS reduces diffusion transformer (DiT) step latency by up to 19.6% relative to tuned online softmax without measurable quality loss. Controlled TPU adaptations of prior GPU stabilization methods show performance or numerical-safety limitations. Preliminary causal LLM experiments also show up to 13.5% higher training throughput in continued pretraining and 7.7% in supervised fine-tuning.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.00444 — 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 2026-11-03T18:12:54.421Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).