Partial Parameter Updates for Efficient Distributed Training

ICLR 2026 Conference Submission20054 Authors

Keywords: distributed training, low-communication, language modeling

TL;DR: Memory and compute efficient method for low-communication distributed training

Abstract: We introduce a memory- and compute-efficient method for low-communication distributed training. Existing methods reduce communication by performing multiple local updates between infrequent global synchronizations. We demonstrate that their efficiency can be significantly improved by restricting backpropagation: instead of updating all the parameters, each node updates only a fixed subset while keeping the remainder frozen during local steps. This constraint substantially reduces peak memory usage and training FLOPs, while a full forward pass over all parameters eliminates the need for cross-node activation exchange. Experiments on a $1.3$B-parameter language model trained across $32$ nodes show that our method matches the perplexity of prior low-communication approaches under identical token and bandwidth budgets while reducing training FLOPs by $15$% and peak memory by up to $47$%.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 20054