English
Related papers

Related papers: APOLLO: SGD-like Memory, AdamW-level Performance

200 papers

Low-rank training has emerged as a promising approach for reducing memory usage in training Large Language Models (LLMs). Previous methods either rely on decomposing weight matrices (e.g., LoRA), or seek to decompose gradient matrices…

Machine Learning · Computer Science 2025-11-18 Xi Chen , Kaituo Feng , Changsheng Li , Xunhao Lai , Xiangyu Yue , Ye Yuan , Guoren Wang

In recent years, Large Language Models (LLMs) have exhibited remarkable capabilities, driving advancements in real-world applications. However, training LLMs on increasingly long input sequences imposes significant challenges due to high…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-03-14 Qiaoling Chen , Shenggui Li , Wei Gao , Peng Sun , Yonggang Wen , Tianwei Zhang

Full fine-tuning of Large Language Models (LLMs) is notoriously memory-intensive, primarily because conventional optimizers such as SGD or Adam assume access to exact gradients derived from cached activations. Existing solutions either…

Machine Learning · Computer Science 2025-10-28 Jing Yang , Kaitong Cai , Yijia Fan , Yufeng Yang , Keze Wang

Fueled by their remarkable ability to tackle diverse tasks across multiple domains, large language models (LLMs) have grown at an unprecedented rate, with some recent models containing trillions of parameters. This growth is accompanied by…

Machine Learning · Computer Science 2025-05-30 Athanasios Glentis , Jiaxiang Li , Qiulin Shang , Andi Han , Ioannis Tsaknakis , Quan Wei , Mingyi Hong

Adaptive optimizers such as Adam (Kingma & Ba, 2015) have been central to the success of large language models. However, they often require to maintain optimizer states throughout training, which can result in memory requirements several…

Machine Learning · Computer Science 2025-02-24 Chao Ma , Wenbo Gong , Meyer Scetbon , Edward Meeds

Large language models (LLMs) have demonstrated remarkable performance due to their large parameter counts and extensive training data. However, their scale leads to significant memory bottlenecks during training, especially when using…

Machine Learning · Computer Science 2026-05-14 Ziqing Wen , Jiahuan Wang , Ping Luo , Dongsheng Li , Tao Sun

Running out of GPU memory has become a main bottleneck for large-scale DNN training. How to reduce the memory footprint during training has received intensive research attention. We find that previous gradient accumulation reduces…

Machine Learning · Computer Science 2023-06-01 Yijia Zhang , Yibo Han , Shijie Cao , Guohao Dai , Youshan Miao , Ting Cao , Fan Yang , Ningyi Xu

The machine learning community has witnessed impressive advancements since large language models (LLMs) first appeared. Yet, their massive memory consumption has become a significant roadblock to large-scale training. For instance, a 7B…

Machine Learning · Computer Science 2024-12-30 Rui Pan , Xiang Liu , Shizhe Diao , Renjie Pi , Jipeng Zhang , Chi Han , Tong Zhang

Fine-tuning large language models (LLMs) often faces GPU memory bottlenecks: the backward pass of first-order optimizers like Adam increases memory usage to more than 10 times the inference level (e.g., 633 GB for OPT-30B). Zeroth-order…

Machine Learning · Computer Science 2025-07-01 Sizhe Dang , Yangyang Guo , Yanjun Zhao , Haishan Ye , Xiaodong Zheng , Guang Dai , Ivor Tsang

Low-Rank Adaptation (LoRA) fine-tunes large models by learning low-rank updates on top of frozen weights, dramatically reducing trainable parameters and memory. However, there is still a gap between full training with low-rank projections…

We introduce AlphaGrad, a memory-efficient, conditionally stateless optimizer addressing the memory overhead and hyperparameter complexity of adaptive methods like Adam. AlphaGrad enforces scale invariance via tensor-wise L2 gradient…

Machine Learning · Computer Science 2025-04-24 Soham Sane

Large Language Models (LLMs) have demonstrated exceptional capabilities across diverse natural language processing benchmarks. However, the escalating scale of model parameters imposes prohibitive memory overheads during training,…

Machine Learning · Computer Science 2026-04-28 Ziqing Wen , Ping Luo , Jiahuan Wang , Kun Yuan , Dongsheng Li , Tao Sun

Training LLMs presents significant memory challenges due to growing size of data, weights, and optimizer states. Techniques such as data and model parallelism, gradient checkpointing, and offloading strategies address this issue but are…

Machine Learning · Computer Science 2024-10-22 Arijit Das

We introduce LDAdam, a memory-efficient optimizer for training large models, that performs adaptive optimization steps within lower dimensional subspaces, while consistently exploring the full parameter space during training. This strategy…

Machine Learning · Computer Science 2025-03-04 Thomas Robert , Mher Safaryan , Ionut-Vlad Modoranu , Dan Alistarh

The AdamW optimizer, while standard for LLM pretraining, is a critical memory bottleneck, consuming optimizer states equivalent to twice the model's size. Although light-state optimizers like SinkGD attempt to address this issue, we…

Machine Learning · Computer Science 2026-04-17 Wooin Lee , Hyun-Tae Kim

It is widely believed that stochastic gradient descent (SGD) performs significantly worse than adaptive optimizers such as Adam in pre-training Large Language Models (LLMs). Yet the underlying reason for this gap remains unclear. In this…

Machine Learning · Computer Science 2026-05-19 Athanasios Glentis , Dawei Li , Chung-Yiu Yau , Mingyi Hong

The vast majority of modern deep learning models are trained with momentum-based first-order optimizers. The momentum term governs the optimizer's memory by determining how much each past gradient contributes to the current convergence…

Machine Learning · Computer Science 2026-05-12 Kristi Topollai , Anna Choromanska

Fine-tuning LLMs is necessary for various dedicated downstream tasks, but classic backpropagation-based fine-tuning methods require substantial GPU memory. To this end, a recent work, MeZO, which relies solely on forward passes to fine-tune…

Machine Learning · Computer Science 2026-05-04 Zhijie Cai , Haolong Chen , Guangxu Zhu

The impressive performance of large language models (LLMs) arises from their massive scale and heterogeneous module composition. However, this structural heterogeneity introduces additional optimization challenges. While adaptive optimizers…

Machine Learning · Computer Science 2026-05-08 Ziqing Wen , Zhouyang Liu , Jiahuan Wang , Ping Luo , Li Shen , Dongsheng Li , Tao Sun

Fine-tuning language models (LMs) with the Adam optimizer often demands excessive memory, limiting accessibility. The "in-place" version of Stochastic Gradient Descent (IP-SGD) and Memory-Efficient Zeroth-order Optimizer (MeZO) have been…

Machine Learning · Computer Science 2024-10-10 Zeman Li , Xinwei Zhang , Peilin Zhong , Yuan Deng , Meisam Razaviyayn , Vahab Mirrokni