https://arxiv.org/abs/2412.11768 Change to arXiv's privacy policy The arXiv Privacy Policy has changed. By continuing to use arxiv.org, you are agreeing to the privacy policy. I Understand Skip to main content Cornell University We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate arxiv logo > cs > arXiv:2412.11768 [ ] Help | Advanced Search [All fields ] Search arXiv logo Cornell University Logo [ ] GO quick links * Login * Help Pages * About Computer Science > Machine Learning arXiv:2412.11768 (cs) [Submitted on 16 Dec 2024 (v1), last revised 17 Dec 2024 (this version, v2)] Title:No More Adam: Learning Rate Scaling at Initialization is All You Need Authors:Minghao Xu, Lichuan Xiang, Xu Cai, Hongkai Wen View a PDF of the paper titled No More Adam: Learning Rate Scaling at Initialization is All You Need, by Minghao Xu and 3 other authors View PDF Abstract:In this work, we question the necessity of adaptive gradient methods for training deep neural networks. SGD-SaI is a simple yet effective enhancement to stochastic gradient descent with momentum (SGDM). SGD-SaI performs learning rate Scaling at Initialization (SaI) to distinct parameter groups, guided by their respective gradient signal-to-noise ratios (g-SNR). By adjusting learning rates without relying on adaptive second-order momentum, SGD-SaI helps prevent training imbalances from the very first iteration and cuts the optimizer's memory usage by half compared to AdamW. Despite its simplicity and efficiency, SGD-SaI consistently matches or outperforms AdamW in training a variety of Transformer-based tasks, effectively overcoming a long-standing challenge of using SGD for training Transformers. SGD-SaI excels in ImageNet-1K classification with Vision Transformers(ViT) and GPT-2 pretraining for large language models (LLMs, transformer decoder-only), demonstrating robustness to hyperparameter variations and practicality for diverse applications. We further tested its robustness on tasks like LoRA fine-tuning for LLMs and diffusion models, where it consistently outperforms state-of-the-art optimizers. From a memory efficiency perspective, SGD-SaI achieves substantial memory savings for optimizer states, reducing memory usage by 5.93 GB for GPT-2 (1.5B parameters) and 25.15 GB for Llama2-7B compared to AdamW in full-precision training settings. Comments: 20 pages, 10 figures Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2412.11768 [cs.LG] (or arXiv:2412.11768v2 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2412.11768 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Minghao Xu [view email] [v1] Mon, 16 Dec 2024 13:41:37 UTC (1,216 KB) [v2] Tue, 17 Dec 2024 09:30:44 UTC (1,216 KB) Full-text links: Access Paper: View a PDF of the paper titled No More Adam: Learning Rate Scaling at Initialization is All You Need, by Minghao Xu and 3 other authors * View PDF * TeX Source * Other Formats license icon view license Current browse context: cs.LG < prev | next > new | recent | 2024-12 Change to browse by: cs cs.AI References & Citations * NASA ADS * Google Scholar * Semantic Scholar a export BibTeX citation Loading... BibTeX formatted citation x [loading... ] Data provided by: Bookmark BibSonomy logo Reddit logo (*) Bibliographic Tools Bibliographic and Citation Tools [ ] Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) [ ] Connected Papers Toggle Connected Papers (What is Connected Papers?) [ ] Litmaps Toggle Litmaps (What is Litmaps?) [ ] scite.ai Toggle scite Smart Citations (What are Smart Citations?) ( ) Code, Data, Media Code, Data and Media Associated with this Article [ ] alphaXiv Toggle alphaXiv (What is alphaXiv?) [ ] Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) [ ] DagsHub Toggle DagsHub (What is DagsHub?) [ ] GotitPub Toggle Gotit.pub (What is GotitPub?) [ ] Huggingface Toggle Hugging Face (What is Huggingface?) [ ] Links to Code Toggle Papers with Code (What is Papers with Code?) [ ] ScienceCast Toggle ScienceCast (What is ScienceCast?) ( ) Demos Demos [ ] Replicate Toggle Replicate (What is Replicate?) [ ] Spaces Toggle Hugging Face Spaces (What is Spaces?) [ ] Spaces Toggle TXYZ.AI (What is TXYZ.AI?) ( ) Related Papers Recommenders and Search Tools [ ] Link to Influence Flower Influence Flower (What are Influence Flowers?) [ ] Core recommender toggle CORE Recommender (What is CORE?) [ ] IArxiv recommender toggle IArxiv Recommender (What is IArxiv?) * Author * Venue * Institution * Topic ( ) About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) * About * Help * Click here to contact arXiv Contact * Click here to subscribe Subscribe * Copyright * Privacy Policy * Web Accessibility Assistance * arXiv Operational Status Get status notifications via email or slack