English
Related papers

Related papers: LoCo: Low-Bit Communication Adaptor for Large-scal…

200 papers

Scalable training of large models (like BERT and GPT-3) requires careful optimization rooted in model design, architecture, and system capabilities. From a system standpoint, communication has become a major bottleneck, especially on…

Machine Learning · Computer Science 2021-07-01 Hanlin Tang , Shaoduo Gan , Ammar Ahmad Awan , Samyam Rajbhandari , Conglong Li , Xiangru Lian , Ji Liu , Ce Zhang , Yuxiong He

Training LLMs relies on distributed implementations using multiple GPUs to compute gradients in parallel with sharded optimizers. However, synchronizing gradients in data parallel setups introduces communication overhead that grows with the…

Machine Learning · Computer Science 2025-10-15 Adel Nabli , Louis Fournier , Pierre Erbacher , Louis Serrano , Eugene Belilovsky , Edouard Oyallon

1-bit gradient compression and local steps are two representative techniques that enable drastic communication reduction in distributed SGD. Their benefits, however, remain an open question on Adam-based large model pre-training (e.g. BERT…

Machine Learning · Computer Science 2022-05-24 Yucheng Lu , Conglong Li , Minjia Zhang , Christopher De Sa , Yuxiong He

To train large models (like BERT and GPT-3) on hundreds of GPUs, communication has become a major bottleneck, especially on commodity systems with limited-bandwidth TCP network. On one side large batch-size optimization such as LAMB…

Machine Learning · Computer Science 2021-10-07 Conglong Li , Ammar Ahmad Awan , Hanlin Tang , Samyam Rajbhandari , Yuxiong He

Training large language models is generally done via optimization methods on clusters containing tens of thousands of accelerators, communicating over a high-bandwidth interconnect. Scaling up these clusters is expensive and can become…

Machine Learning · Computer Science 2025-06-13 Jari Kolehmainen , Nikolay Blagoev , John Donaghy , Oğuzhan Ersoy , Christopher Nies

Communication-efficient distributed training algorithms have received considerable interest recently due to their benefits for training Large Language Models (LLMs) in bandwidth-constrained settings, such as across datacenters and over the…

Machine Learning · Computer Science 2025-11-07 Amir Sarfi , Benjamin Thérien , Joel Lidin , Eugene Belilovsky

In the decentralized distributed learning, achieving fast convergence and low communication cost is essential for scalability and high efficiency. Adaptive gradient methods, such as Adam, have demonstrated strong practical performance in…

Machine Learning · Computer Science 2026-04-14 Wei Liu , Anweshit Panda , Ujwal Pandey , Haven Cook , George M. Slota , Naigang Wang , Jie Chen , Yangyang Xu

Distributed model training suffers from communication bottlenecks due to frequent model updates transmitted across compute nodes. To alleviate these bottlenecks, practitioners use gradient compression techniques like sparsification,…

Machine Learning · Computer Science 2020-11-02 Saurabh Agarwal , Hongyi Wang , Kangwook Lee , Shivaram Venkataraman , Dimitris Papailiopoulos

In Distributed optimization and Learning, and even more in the modern framework of federated learning, communication, which is slow and costly, is critical. We introduce LoCoDL, a communication-efficient algorithm that leverages the two…

Optimization and Control · Mathematics 2025-03-03 Laurent Condat , Artavazd Maranjyan , Peter Richtárik

Large Language Models (LLMs) have revolutionized Natural Language Processing (NLP) but demand massive GPU resources for training. Lowering the threshold for LLMs training would encourage greater participation from researchers, benefiting…

Computation and Language · Computer Science 2024-06-07 Kai Lv , Yuqing Yang , Tengxiao Liu , Qinghui Gao , Qipeng Guo , Xipeng Qiu

Despite advances using low-rank adapters and quantization, pretraining of large models on consumer hardware has not been possible without model sharding, offloading during training, or per-layer gradient updates. To address these…

Machine Learning · Computer Science 2024-11-05 Sebastian Loeschcke , Mads Toftrup , Michael J. Kastoryano , Serge Belongie , Vésteinn Snæbjarnarson

Large language models have achieved remarkable success, but their extensive parameter size necessitates substantial memory for training, thereby setting a high threshold. While the recently proposed low-memory optimization (LOMO) reduces…

Machine Learning · Computer Science 2024-06-07 Kai Lv , Hang Yan , Qipeng Guo , Haijun Lv , Xipeng Qiu

The training of large models demands substantial computational resources, typically available only in data centers with high-bandwidth interconnects. However, reducing the reliance on high-bandwidth interconnects between nodes enables the…

Machine Learning · Computer Science 2025-10-07 Sasho Nedelkoski , Alexander Acker , Odej Kao , Soeren Becker , Dominik Scheinert

Handling communication overhead in large-scale tensor-parallel training remains a critical challenge due to the dense, near-zero distributions of intermediate tensors, which exacerbate errors under frequent communication and introduce…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-04-28 Man Liu , Xingchen Liu , Xingjian Tian , Bing Lu , Shengkay Lyu , Shengquan Yin , Wenjing Huang , Zheng Wei , Hairui Zhao , Guangming Tan , Dingwen Tao

Training deep neural networks on large datasets containing high-dimensional data requires a large amount of computation. A solution to this problem is data-parallel distributed training, where a model is replicated into several…

Machine Learning · Computer Science 2021-03-18 Lusine Abrahamyan , Yiming Chen , Giannis Bekoulis , Nikos Deligiannis

Communication compression is a crucial technique for modern distributed learning systems to alleviate their communication bottlenecks over slower networks. Despite recent intensive studies of gradient compression for data parallel-style…

Machine Learning · Computer Science 2023-03-08 Jue Wang , Binhang Yuan , Luka Rimanic , Yongjun He , Tri Dao , Beidi Chen , Christopher Re , Ce Zhang

Recent research explores optimization using large language models (LLMs) by either iteratively seeking next-step solutions from LLMs or directly prompting LLMs for an optimizer. However, these approaches exhibit inherent limitations,…

Optimization and Control · Mathematics 2024-03-06 Zeyuan Ma , Hongshu Guo , Jiacheng Chen , Guojun Peng , Zhiguang Cao , Yining Ma , Yue-Jiao Gong

Training large models with distributed data parallelism (DDP) requires frequent communication of gradients across workers, which can saturate bandwidth. Infrequent communication strategies (e.g., Local SGD) reduce this overhead but, when…

Large language models (LLM) have become a critical component in many applications of machine learning. However, standard approaches to training LLM require a large number of tightly interconnected accelerators, with devices exchanging…

Scaling neural network training increasingly depends on synchronous data-parallelism, yet full-precision gradient all-reduce imposes a severe communication bottleneck. We propose Decoupled Momentum Optimization (DeMo), a drop-in replacement…

Machine Learning · Computer Science 2026-02-10 Bowen Peng , Lizhang Chen , Baiyu Su , Jeffrey Quesnelle , Diederik P. Kingma , Qiang Liu
‹ Prev 1 2 3 10 Next ›