블로그 글 추가: 2026-06-03-using-muon-optimizer-with-deepspeed, DeepSpeed에서 Muon 옵티마이저 사용하기#91
Open
9bow wants to merge 2 commits into
Open
블로그 글 추가: 2026-06-03-using-muon-optimizer-with-deepspeed, DeepSpeed에서 Muon 옵티마이저 사용하기#919bow wants to merge 2 commits into
9bow wants to merge 2 commits into
Conversation
PR Preview빌드가 완료되었습니다! 아래 링크에서 변경사항을 확인할 수 있습니다. 미리보기: https://pytorchkr-pr-preview-91.surge.sh
|
- frontier AI Labs: '최전선 AI 연구소들' → '선도적인 AI 연구소들' - how it performs: '어떻게 동작하는지' → '어떤 성능을 보이는지' (performance 뉘앙스 반영)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
번역 글 소개
DeepSpeed에서 Muon 옵티마이저 사용하기 번역 글을 추가합니다.
이제 DeepSpeed가 Muon 옵티마이저를 지원합니다. Muon은 신경망의 은닉 2D 가중치를 위해 설계된 옵티마이저로, 모멘텀 행렬을 Newton-Schulz 반복으로 직교화하여 사전 학습에서의 수렴 이점을 얻으며, 모멘텀 버퍼를 하나만 유지해 Adam보다 옵티마이저 상태 메모리를 절약합니다. 이번 글은 DeepSpeed에서의 Muon 지원 방식과 함께, Moonlight-16B-A3B 미세 조정 수렴 실험(4개 지표 중 3개에서 AdamW 상회) 및 Qwen2.5-3B 메모리 측정 결과(GPU당 약 9% 절감)를 소개합니다.