AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

LiMuon: Light and Fast Muon Optimizer for Large Models

Feihu Huang, Yuning Luo, Songcan Chen

ICML 2026 · Optimization › Nonconvex and smooth optimization · approach: Complexity of nonconvex first-order methods · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase LiMuon rank from 4 to 16 (ablation)LiMuon · Mamba-130M WikiText-103validation perplexity: 79.12 -> 53.84Improved
use full rank instead of rank 16 (ablation)LiMuon · Mamba-130M WikiText-103validation perplexity: 53.84 -> 47.78Improved
increase LiMuon rank from 4 to full rank (ablation)LiMuon · Mamba-130Mmemory (GB): 18.86 -> 22.80Hurt
increase LiMuon rank from 4 to 8 (ablation)LiMuon · Qwen2.5-0.5B MiniPilevalidation perplexity: 92.06 -> 54.87Improved
use full rank instead of rank 16 (ablation)LiMuon · Qwen2.5-0.5B MiniPilevalidation perplexity: 46.77 -> 40.83Improved
use LiMuon rank=8 instead of Muonoptimizer · Qwen2.5-0.5Bmemory (GB): 54.14 -> 52.55Improved
use LiMuon full rank instead of Muonoptimizer · Qwen2.5-0.5Bvalidation perplexity: 67.60 -> 40.83Improved
use LiMuon full rank instead of Muonoptimizer · ViT Tiny-ImageNetvalidation top-1 accuracy (%): 47.87 -> 48.04Improved
use LiMuon rank 4 instead of Muonoptimizer · ViT Tiny-ImageNettraining top-1 accuracy (%): 78.96 -> 64.73Hurt
use LiMuon full rank instead of Muonoptimizer · ViTtime to target: 7min 6s -> 7min 1sImproved
use low-rank LiMuon (rank 4) instead of Muonoptimizer · ViTtime to target: 7min 6s -> 12min 4sHurt
use LiMuon rank 4 instead of Muonoptimizer · Mamba-130Mvalidation perplexity: 71.27 -> 79.12Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy