AI Paper Atlas

Research map / Optimization / Optimizers

Memory-Efficient LLM Pretraining via Minimalist Optimizer Design

Athanasios Glentis, Jiaxiang Li, Andi Han, Mingyi Hong

ICML 2026 · Optimization › Optimizers · approach: Memory-efficient LLM optimizers · 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
add last-layer momentum to column-wise SGD (ablation)SGD column-wise · C4 LLaMA-130Mperplexity: 28.85 -> 22.57Improved
add last-layer momentum to column-wise SGD (ablation)SGD column-wise · C4 LLaMA-60Mperplexity: 39.89 -> 30.81Improved
add last-layer momentum to column-wise SGD (ablation)SGD column-wise · C4 LLaMA-350Mperplexity: 20.38 -> 16.32Improved
add first-layer momentum on top of last-layer momentum (ablation)SCALE · C4 LLaMA-130Mperplexity: 22.57 -> 22.26No clear effect
column-wise instead of row-wise normalization (ablation)SGD · C4 LLaMA-130Mperplexity: 37.67 -> 28.85Improved
column-wise instead of sign normalization (ablation)SGD · C4 LLaMA-130Mperplexity: 40.42 -> 28.85Improved
singular-value (NS) instead of column-wise normalization, no momentum (ablation)SGD · C4 LLaMA-130Mperplexity: 25.25 vs 28.85Hurt
column-wise instead of row-wise normalization on last layer (ablation)SCALE · C4 LLaMA-130Mperplexity: 28.83 -> 22.57Improved
column-wise instead of row-wise normalization on non-last layers (ablation)SCALE · C4 LLaMA-130Mperplexity: 23.27 -> 22.57Improved
column-wise instead of row-wise normalization on first layer (ablation)SCALE · C4 LLaMA-130Mperplexity: 22.94 -> 22.57Improved
column-wise instead of normalizing along larger dimension (ablation)SCALE · C4 LLaMA-130Mperplexity: 23.52 -> 22.57Improved
singular-value (NS) instead of column-wise with last-layer momentum (ablation)SCALE · C4 LLaMA-350Mperplexity: 16.67 vs 16.32Hurt

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