Metraipolis › archive › dc_d811ec4df67e
G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation — arXiv:2609.31009
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.31009
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T01:39:39.862Z
- SHA-256 of the archived text
fbf3689bd7d05ecd4d8dc9ea141152f82eca4f10c1b57acc0fe0727d55765c7f- Stable URL of this page
https://metraipolis.com/documento/dc_d811ec4df67e
How to cite
G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation — arXiv:2609.31009 — arXiv (en), https://arxiv.org/abs/2609.31009 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d811ec4df67e (Metraipolis archive dc_d811ec4df67e, sha256 fbf3689bd7d05ecd4d8dc9ea141152f82eca4f10c1b57acc0fe0727d55765c7f)
BibTeX
@misc{dcd811ec4df67e,
title = {G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation — arXiv:2609.31009},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.31009},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d811ec4df67e; sha256 fbf3689bd7d05ecd4d8dc9ea141152f82eca4f10c1b57acc0fe0727d55765c7f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31009. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Post-training quantization (PTQ) is a practical approach to reducing the memory and computational footprint of large language models (LLMs) without retraining. GPTQ-based methods have become the de facto standard, yet they suffer from two complementary limitations. Methods with local, layer-wise objectives lack global supervision; while methods with global objectives fix their Hessian estimates at the start and ignore first-order gradients, so their guidance grows stale as quantization proceeds. This paper presents G$^2$PTQ, a unified PTQ framework with Generalized Gradient Compensation that integrates both first- and second-order information under a globally supervised, block-wise optimization objective. By refreshing gradient and Hessian estimates before quantizing each Transformer block, G$^2$PTQ avoids the staleness of prior global methods. Furthermore, to stabilize the exact first-order compensation, we introduce a trust-region scaling mechanism that dynamically bounds the gradient step to prevent exploding weight updates. Finally, we derive efficient implementations for block-wise Hessian approximation and exact gradient compensation. Experimental results on various model families and bit-widths demonstrate that G$^2$PTQ enables better alignment with the full-precision model, outperforming state-of-the-art baselines. Code is available at: this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.31009 — 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 2026-10-08T01:12:14.524Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).