Federated Full-Parameter Tuning of Billion-Sized Language Models with Communication Cost under 18 Kilobytes

Zhen Qin, Daoyuan Chen, Bingchen Qian, Bolin Ding, Yaliang Li, Shuiguang Deng
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:41473-41497, 2024.

Abstract

Pre-trained large language models (LLMs) need fine-tuning to improve their responsiveness to natural language instructions. Federated learning offers a way to fine-tune LLMs using the abundant data on end devices without compromising data privacy. Most existing federated fine-tuning methods for LLMs rely on parameter-efficient fine-tuning techniques, which may not reach the performance height possible with full-parameter tuning. However, federated full-parameter tuning of LLMs is a non-trivial problem due to the immense communication cost. This work introduces FedKSeed that employs zeroth-order optimization with a finite set of random seeds. It significantly reduces transmission requirements between the server and clients to just a few random seeds and scalar gradients, amounting to only a few thousand bytes, making federated full-parameter tuning of billion-sized LLMs possible on devices. Building on it, we develop a strategy enabling probability-differentiated seed sampling, prioritizing perturbations with greater impact on model accuracy. Experiments across six scenarios with various LLMs, datasets and data partitions demonstrate that our approach outperforms existing federated LLM fine-tuning methods in both communication efficiency and zero-shot generalization.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-qin24a, title = {Federated Full-Parameter Tuning of Billion-Sized Language Models with Communication Cost under 18 Kilobytes}, author = {Qin, Zhen and Chen, Daoyuan and Qian, Bingchen and Ding, Bolin and Li, Yaliang and Deng, Shuiguang}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {41473--41497}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/qin24a/qin24a.pdf}, url = {https://proceedings.mlr.press/v235/qin24a.html}, abstract = {Pre-trained large language models (LLMs) need fine-tuning to improve their responsiveness to natural language instructions. Federated learning offers a way to fine-tune LLMs using the abundant data on end devices without compromising data privacy. Most existing federated fine-tuning methods for LLMs rely on parameter-efficient fine-tuning techniques, which may not reach the performance height possible with full-parameter tuning. However, federated full-parameter tuning of LLMs is a non-trivial problem due to the immense communication cost. This work introduces FedKSeed that employs zeroth-order optimization with a finite set of random seeds. It significantly reduces transmission requirements between the server and clients to just a few random seeds and scalar gradients, amounting to only a few thousand bytes, making federated full-parameter tuning of billion-sized LLMs possible on devices. Building on it, we develop a strategy enabling probability-differentiated seed sampling, prioritizing perturbations with greater impact on model accuracy. Experiments across six scenarios with various LLMs, datasets and data partitions demonstrate that our approach outperforms existing federated LLM fine-tuning methods in both communication efficiency and zero-shot generalization.} }
Endnote
%0 Conference Paper %T Federated Full-Parameter Tuning of Billion-Sized Language Models with Communication Cost under 18 Kilobytes %A Zhen Qin %A Daoyuan Chen %A Bingchen Qian %A Bolin Ding %A Yaliang Li %A Shuiguang Deng %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-qin24a %I PMLR %P 41473--41497 %U https://proceedings.mlr.press/v235/qin24a.html %V 235 %X Pre-trained large language models (LLMs) need fine-tuning to improve their responsiveness to natural language instructions. Federated learning offers a way to fine-tune LLMs using the abundant data on end devices without compromising data privacy. Most existing federated fine-tuning methods for LLMs rely on parameter-efficient fine-tuning techniques, which may not reach the performance height possible with full-parameter tuning. However, federated full-parameter tuning of LLMs is a non-trivial problem due to the immense communication cost. This work introduces FedKSeed that employs zeroth-order optimization with a finite set of random seeds. It significantly reduces transmission requirements between the server and clients to just a few random seeds and scalar gradients, amounting to only a few thousand bytes, making federated full-parameter tuning of billion-sized LLMs possible on devices. Building on it, we develop a strategy enabling probability-differentiated seed sampling, prioritizing perturbations with greater impact on model accuracy. Experiments across six scenarios with various LLMs, datasets and data partitions demonstrate that our approach outperforms existing federated LLM fine-tuning methods in both communication efficiency and zero-shot generalization.
APA
Qin, Z., Chen, D., Qian, B., Ding, B., Li, Y. & Deng, S.. (2024). Federated Full-Parameter Tuning of Billion-Sized Language Models with Communication Cost under 18 Kilobytes. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:41473-41497 Available from https://proceedings.mlr.press/v235/qin24a.html.

Related Material