paper-with-me

Papers

BaKron: Efficient Quantization with Kronecker-Factored Hessians

2026-08-06 · Johann Birnick, Rayan Saab arxiv

We accelerate a family of algorithms for neural network quantization whose geometry is informed by any Kronecker-factored approximation of the Hessian. GPTQ-style adaptive rounding typically uses one-sided information derived from input activations. Two-sided Kronecker-factored Hessian approximations can additionally capture correlations across output coordinates, but applying GPTQ directly in the vectorized weight domain is computationally expensive. Building on the two-sided adaptive-rounding formulation used by BoA and YAQA, we introduce BaKron, an efficient solver that combines anti-diagonal parallelism with a recursive divide-and-conquer construction. For an $m\times n$ weight matrix, BaKron uses $O(m+n)$ sequential steps while reducing the total work from $O(m^2n^2)$ to $O(mn(m+n))$. Thus, it matches the cubic scaling of GPTQ while exploiting richer curvature information. Moreover, BaKron is modular with respect to both the base quantizer and the Hessian estimator. We also provide practical benchmarks, consider a range of Hessians that BaKron can be called with, find an efficient technique to compute these Hessians, and evaluate the algorithm experimentally.

📄 PDF Abstract BibTeX arXiv:2608.06291

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

KronQ: LLM Quantization via Kronecker-Factored Hessian

2026-07-08 · Donghyun Lee, Yuhang Li, Ruokai Yin, Priyadarshini Panda arxiv

Post-training quantization (PTQ) is a widely adopted technique for compressing large language models (LLMs) without retraining. Existing second-order PTQ methods, including GPTQ, construct quantization objectives exclusi…

Model-Preserving Adaptive Rounding

2025-05-29 · Albert Tseng, Zhaofeng Sun, Christopher De Sa

The main goal of post-training quantization (PTQ) is to produced a compressed model whose output distribution is as close to the original model's as possible. To do this tractably, almost all LLM PTQ algorithms quantize …

modelQuantization

Rethinking Bregman Divergences in Kronecker-Factored Optimizers

2026-05-30 · Bing Liu, Wenjie Zhou, Chengcheng Zhao arxiv

Shampoo-style optimizers approximate gradient covariance matrices using Kronecker-factored structures. Recent work~\cite{lin2026understanding} showed that such approximations can be viewed as projections under Bregman ma…

Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation

2017-08-17 · NeurIPS 2017 12 · Yuhuai Wu, Elman Mansimov, Shun Liao, Roger Grosse 외

In this work, we propose to apply trust region optimization to deep reinforcement learning using a recently proposed Kronecker-factored approximation to the curvature. We extend the framework of natural policy gradient a…

Atari Gamescontinuous-controlContinuous ControlDeep Reinforcement Learning+4

Eigenvalue-corrected Natural Gradient Based on a New Approximation

2020-11-27 · Kai-Xin Gao, Xiao-Lei Liu, Zheng-Hai Huang, Min Wang 외

Using second-order optimization methods for training deep neural networks (DNNs) has attracted many researchers. A recently proposed method, Eigenvalue-corrected Kronecker Factorization (EKFAC) (George et al., 2018), pro…