English(EN) Scale Weight Decay and Train Better

新的缩放权重衰减方法加速神经网络训练

作者 PulseAugur 编辑部 · [1 个来源] · 2026-07-28 04:00

研究人员引入了一种新颖的缩放权重衰减方法，该方法受Robbins-Monro条件的启发，以改进神经网络训练。该技术根据峰值学习率的比例调整权重衰减，确保渐近平稳性，并避免恒定权重衰减引入的偏差。当应用于使用Muon优化器的混合专家模型时，这种缩放权重衰减（Muon-SW）显示出显著的加速效果，在大规模训练中，达到目标验证损失的速度最多可提高30%。 AI

影响有望以最小的实施工作量显著加速前沿模型的预训练。

排序理由该集群包含一篇详细介绍神经网络新训练方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍神经网络新训练方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, infra; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 102 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Anuj Apte · 2026-07-28 04:00

缩放权重衰减并进行更好的训练

arXiv:2607.23777v1 Announce Type: cross Abstract: The discovery of scaling laws has motivated training neural networks on ever increasing quantities of data. This is typically done with a constant decoupled weight decay which causes the network weights to shrink steadily over the…

本文如何被排名

报道来源 [1]

缩放权重衰减并进行更好的训练

相关实体

相关话题