Mitigating Training Imbalance in LLM Fine-Tuning via Selective Parameter Merging
Yiming Ju, Ziyi Ni, Xingrun Xing, Zhixiong Zeng, Hanyu Zhao, Siqi Fan, Zheng Zhang
Abstract
Supervised fine-tuning (SFT) is crucial for adapting Large Language Models (LLMs) to specific tasks. In this work, we demonstrate that the order of training data can lead to significant training imbalances, potentially resulting in performance degradation. Consequently, we propose to mitigate this imbalance by merging SFT models fine-tuned with different data orders, thereby enhancing the overall effectiveness of SFT. Additionally, we introduce a novel technique, “parameter-selection merging,” which outperforms traditional weighted-average methods on five datasets. Further, through analysis and ablation studies, we validate the effectiveness of our method and identify the sources of performance improvements.
BibTeX
@inproceedings{ju-etal-2024-mitigating,
title = "Mitigating Training Imbalance in {LLM} Fine-Tuning via Selective Parameter Merging",
author = "Ju, Yiming and
Ni, Ziyi and
Xing, Xingrun and
Zeng, Zhixiong and
Zhao, Hanyu and
Fan, Siqi and
Zhang, Zheng",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.892/",
doi = "10.18653/v1/2024.emnlp-main.892",
pages = "15952--15959"
}