1
PC Layer: Polynomial Weight Preconditioning for Improving LLM Pre-Training
多项式预条件层PC Layer通过重塑权重矩阵奇异值谱,稳定大模型预训练过程,提升训练质量与收敛效率。
arXiv:2606.06470v1 Announce Type: cross Abstract: We propose a preconditioning (PC) layer, a weight parameterization via polynomial preconditioner tha…