PulseAugur
实时 06:34:58

新的 Muon 优化器变体提高了语言模型预训练效率

研究人员开发了两种新的 Muon 优化器变体,名为 Muon-NSRMuon-VS,旨在提高语言模型预训练的效率。这些变体通过整合梯度方差信息来调整 Muon 的正交动量更新,类似于 Adam 风格的方法。在 Llama 风格和 GPT-2 模型上的实验表明,特别是 Muon-VS,与标准的 Muon 优化器相比,在达到目标验证损失方面实现了 1.33 倍的加速。 AI

影响 将方差自适应调制引入 Muon 风格的优化器,有可能降低大型语言模型的计算成本并加速训练。

排序理由 该集群包含一篇详细介绍语言模型预训练优化器新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

新的 Muon 优化器变体提高了语言模型预训练效率

本文如何被排名

Signal score
29 / 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语言模型预训练优化器新方法的学术论文。[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
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准

报道来源 [1]

  1. arXiv cs.LG TIER_1 English(EN) · Jingru Li, Yibo Fan, Huan Li ·

    方差自适应μ子:用于高效语言模型预训练的预正交化方差调制

    arXiv:2601.14603v2 Announce Type: replace Abstract: Optimizer design plays a central role in efficient language model pretraining, directly affecting optimization dynamics, convergence speed, and compute cost under fixed training budgets. Muon has emerged as a strong optimizer by…