Metraipolis › archive › dc_2aca4a4b993e

CentriQ: Calibration-Free Quantization of Diffusion Transformers via Exact Mean Centering — arXiv:2610.06260

Source
arXiv
Original
https://arxiv.org/abs/2610.06260
License (as archived)
CC0-1.0
Archived
2026-10-06T05:27:33.044Z
SHA-256 of the archived text
842631d7292529ba61d8e2cb58a8d00f1e41ef6f186863a05c9440000febf4eb
Stable URL of this page
https://metraipolis.com/documento/dc_2aca4a4b993e

How to cite

CentriQ: Calibration-Free Quantization of Diffusion Transformers via Exact Mean Centering — arXiv:2610.06260 — arXiv, https://arxiv.org/abs/2610.06260 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2aca4a4b993e (Metraipolis archive dc_2aca4a4b993e, sha256 842631d7292529ba61d8e2cb58a8d00f1e41ef6f186863a05c9440000febf4eb)
BibTeX
@misc{dc2aca4a4b993e,
  title = {CentriQ: Calibration-Free Quantization of Diffusion Transformers via Exact Mean Centering — arXiv:2610.06260},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06260},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2aca4a4b993e; sha256 842631d7292529ba61d8e2cb58a8d00f1e41ef6f186863a05c9440000febf4eb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06260. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion transformers (DiTs) achieve state-of-the-art image generation, but their sampling cost limits deployment. Quantizing both weights and activations to 4 bits reduces this cost, yet existing methods fall short in one of two ways. Calibration-based methods are tied to a specific checkpoint and prompt distribution, whereas data-free Hadamard rotation, effective for LLMs, loses quality on DiTs. We show that this loss has a structural cause. Adaptive layer-norm conditioning adds a per-token mean to the activations, and at the widths of the evaluated DiTs, the Hadamard rotations used by data-free methods cannot spread this mean uniformly across coordinates. A single dominant direction therefore survives the rotation and sets the quantization range. We introduce CentriQ, a calibration-free quantizer that centers each token before rotation and restores the mean exactly through a rank-1 full-precision branch, so that per-token scales follow in closed form without data. Weights are fitted under a robust $\ell_p$ objective that tracks the dense mode of each group and discounts heavy tails. Across three DiTs, CentriQ matches the quality of calibrated SVDQuant at 4 bits, whereas calibration-free weight quantizers with plain per-token activation quantization collapse or degrade substantially. CentriQ outperforms the strongest calibration-free method reported to date at 2-bit weights. It is also the first calibration-free method to retain usable image quality at 2-bit activations.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI); Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.06260 — 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 2026-10-06T05:24:06.136Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).