arXiv Machine Learning

Probe-Space Preconditioning for Fast and Stable Zero-Order Training

arXiv AI
Oct 15

Clean: Second-order LLM Training at Linear Memory Cost via Nystr{\"o}m Sketching

Clean is a memory‑efficient optimizer that uses Nyström sketching to approximate SOAP’s curvature preconditioners, reducing memory complexity from quadratic to linear while preserving rich curvature information. It also reintegrates off‑subspace components and offers a low‑precision variant, Q‑Clean, that cuts optimizer memory by over 50% compared to Muon while maintaining strong performance. Clean outperforms AdamW in wall‑clock time and enables pre‑training a 13B‑parameter model on a single 80GB GPU.

By Beheshteh T. Rakhshan, Sahar Rajabi, Maziar Sargordi, Shikai Fang, Guillaume Rabusseau, Sirisha Rambhatla
arXiv Machine Learning
Oct 3

TACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning

The paper introduces TACO, a new optimizer for fine‑tuning large language models that drastically reduces optimizer state memory while preserving first‑order gradients. TACO selects the sign of the largest magnitude entry in each column of weight matrices, achieving a 174× reduction in persistent optimizer memory compared to AdamW8bit and a 2.9× decrease in peak training memory on OPT‑13B. This allows full‑parameter fine‑tuning of 30–32B‑parameter models on a single 80 GB GPU across multiple model families and tasks, with comparable accuracy and runtime to existing methods.

By Jichao Jiang (University of Central Florida), Cristian McGee (University of Central Florida), El Houcine Bergou (Mohammed VI Polytechnic University), Hanqin Cai (University of Central Florida), Aritra Dutta (University of Central Florida)
arXiv AI
Jul 24

SOAP, Muon, and Beyond: Pushing LLM Pretraining Scales

arXiv:2607. 20548v1 Announce Type: cross Abstract: Higher-order optimizers such as Muon and SOAP offer faster convergence than AdamW, but their computational cost and numerical stability challenges have limited adoption at scale.

By Mikail Khona, Aditya Vavre, Boxiang Wang, Deyu Fu, Hao Wu, Mike Chrzanowski, Bryan Catanzaro, Dheevatsa Mudigere, Jeff Pool, Michael Lightstone, Mohammad Shoeybi, Mostofa Patwary, Nima Tajbakhsh, Tijmen Blankevoort
arXiv Machine Learning
Oct 12

BLOKS: Scaling Kernel Solvers with Sketched Block Preconditioning

BLOKS is a Conjugate Gradient-based solver for Nyström‑sketched Kernel Ridge Regression that introduces a novel preconditioner combining block Jacobi approximation with sketching. This preconditioner’s memory cost is independent of dataset size and model parameters, enabling scaling to large models while maintaining GPU acceleration. Experiments on standard large‑scale datasets show BLOKS achieves accuracy comparable to or better than state‑of‑the‑art kernel solvers with favorable time/memory complexity.

By Giordano Vitale, Giacomo Meanti, Marco Rando, Blanca Ayuso de Dios, Lorenzo Rosasco
arXiv AI
Jun 15

Gefen: Optimized Stochastic Optimizer

arXiv:2606. 13894v1 Announce Type: cross Abstract: AdamW is a default optimizer for modern deep learning, but its first and second moment states add roughly two parameter-sized buffers to training memory.

By Nadav Benedek, Tomer Koren, Ohad Fried
arXiv AI
Oct 27

Fantastic Pretraining Optimizers and Where to Find Them

The paper evaluates ten deep learning optimizers across models ranging from 0.1B to 1.2B parameters and varying data-to-model ratios, highlighting that fair comparisons require rigorous hyperparameter tuning and end-of-training evaluations. It finds that claimed speedups over AdamW are lower than advertised, dropping to only 1.1× for the largest models, and that intermediate checkpoint comparisons can mislead due to learning‑rate decay. The study shows that matrix‑based optimizers like Muon and Soap achieve the highest speedups, but these benefits diminish as model size increases.

By Kaiyue Wen, David Hall, Tengyu Ma, Percy Liang