Metraipolis › archive › dc_c330a19aa5bf

Post-training Quantization for Hybrid Iterative Generative Models — arXiv:2608.13932

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.13932
License (as archived)
CC0-1.0
Archived
2026-10-19T02:30:27.605Z
SHA-256 of the archived text
793a2edda55f82d00e7e669941c631c30a0825e42c12df408149499698f1ee9f
Stable URL of this page
https://metraipolis.com/documento/dc_c330a19aa5bf

How to cite

Post-training Quantization for Hybrid Iterative Generative Models — arXiv:2608.13932 — arXiv (en), https://arxiv.org/abs/2608.13932 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c330a19aa5bf (Metraipolis archive dc_c330a19aa5bf, sha256 793a2edda55f82d00e7e669941c631c30a0825e42c12df408149499698f1ee9f)
BibTeX
@misc{dcc330a19aa5bf,
  title = {Post-training Quantization for Hybrid Iterative Generative Models — arXiv:2608.13932},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.13932},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c330a19aa5bf; sha256 793a2edda55f82d00e7e669941c631c30a0825e42c12df408149499698f1ee9f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13932. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Iterative Generative Models (IGMs) span autoregressive and diffusion paradigms, and hybrid variants that couple them can achieve remarkable image-generation fidelity. However, their iterative inference incurs substantial computational overhead, making Post-training Quantization (PTQ) appealing for acceleration, while directly applying vanilla PTQ to hybrid IGMs can trigger model collapse. By analyzing these failures, we identify two critical challenges: Excessive Outliers (EOs) in the activations create an irreconcilable trade-off between preserving normal precision and covering EOs, resulting in severe degradation in generation quality; Amplified Anomalies (AAs) arising unpredictably from minor quantization errors, create a mismatch between calibration and inference, thus iteratively triggering model collapse. To address these challenges, we introduce HyGenQ, a PTQ framework for hybrid IGMs. HyGenQ comprises Hierarchical Cluster Decoupling (HCD) and Scaling Recalibration (SR). HCD identifies and decouples outlier channels via a multi-stage clustering process, effectively isolating EOs while maintaining normal value precision, thereby alleviating performance degradation. SR scales AAs beyond Gaussian Bound, thereby avoiding model collapse caused by aggressive truncation. Extensive experiments demonstrate that HyGenQ successfully quantizes representative hybrid IGMs to 8-bit precision (W8A8), significantly outperforming existing baselines and validating its robustness across different model families.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-17 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.13932 — 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 2026-10-19T02:26:17.782Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).