6 papers
Smoothing DiLoCo with Primal Averaging for Faster Training of LLMs
Aaron Defazio, Konstantin Mishchenko, Parameswaran Raman +2
We propose Generalized Primal Averaging (GPA), an extension of Nesterov's method that unifies and generalizes recent averaging-based optimizers like single-worker DiLoCo and Schedu…
Stochastic Approximation with Block Coordinate Optimal Stepsizes
Tao Jiang, Lin Xiao
We consider stochastic approximation with block-coordinate stepsizes and propose adaptive stepsize rules that aim to minimize the expected distance from the next iterate to an (unk…
ParetoQ: Improving Scaling Laws in Extremely Low-bit LLM Quantization
Zechun Liu, Changsheng Zhao, Hanxian Huang +13
The optimal bit-width for achieving the best trade-off between quantized model size and accuracy has been a subject of ongoing debate. While some advocate for 4-bit quantization, o…
Bregman Douglas-Rachford Splitting Method
Shiqian Ma, Lin Xiao, Renbo Zhao
In this paper, we propose the Bregman Douglas-Rachford splitting (BDRS) method and its variant Bregman Peaceman-Rachford splitting method for solving maximal monotone inclusion pro…
Quantization through Piecewise-Affine Regularization: Optimization and Statistical Guarantees
Jianhao Ma, Lin Xiao
Optimization problems over discrete or quantized variables are very challenging in general due to the combinatorial nature of their search space. Piecewise-affine regularization (P…
PARQ: Piecewise-Affine Regularized Quantization
Lisa Jin, Jianhao Ma, Zechun Liu +3
We develop a principled method for quantization-aware training (QAT) of large-scale machine learning models. Specifically, we show that convex, piecewise-affine regularization (PAR…