Parameter Competition Balancing for Model Merging

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Du, Guodong, Lee, Junlin, Li, Jing, Jiang, Runhua, Guo, Yifei, Yu, Shuyang, Liu, Hanting, Goh, Sim Kuan, Tang, Ho-Kin, He, Daojing, Zhang, Min
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912056753520640
author Du, Guodong
Lee, Junlin
Li, Jing
Jiang, Runhua
Guo, Yifei
Yu, Shuyang
Liu, Hanting
Goh, Sim Kuan
Tang, Ho-Kin
He, Daojing
Zhang, Min
author_facet Du, Guodong
Lee, Junlin
Li, Jing
Jiang, Runhua
Guo, Yifei
Yu, Shuyang
Liu, Hanting
Goh, Sim Kuan
Tang, Ho-Kin
He, Daojing
Zhang, Min
contents While fine-tuning pretrained models has become common practice, these models often underperform outside their specific domains. Recently developed model merging techniques enable the direct integration of multiple models, each fine-tuned for distinct tasks, into a single model. This strategy promotes multitasking capabilities without requiring retraining on the original datasets. However, existing methods fall short in addressing potential conflicts and complex correlations between tasks, especially in parameter-level adjustments, posing a challenge in effectively balancing parameter competition across various tasks. This paper introduces an innovative technique named PCB-Merging (Parameter Competition Balancing), a lightweight and training-free technique that adjusts the coefficients of each parameter for effective model merging. PCB-Merging employs intra-balancing to gauge parameter significance within individual tasks and inter-balancing to assess parameter similarities across different tasks. Parameters with low importance scores are dropped, and the remaining ones are rescaled to form the final merged model. We assessed our approach in diverse merging scenarios, including cross-task, cross-domain, and cross-training configurations, as well as out-of-domain generalization. The experimental results reveal that our approach achieves substantial performance enhancements across multiple modalities, domains, model sizes, number of tasks, fine-tuning forms, and large language models, outperforming existing model merging methods. The code is publicly available at: \url{https://github.com/duguodong7/pcb-merging}.
format Preprint
id arxiv_https___arxiv_org_abs_2410_02396
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Parameter Competition Balancing for Model Merging
Du, Guodong
Lee, Junlin
Li, Jing
Jiang, Runhua
Guo, Yifei
Yu, Shuyang
Liu, Hanting
Goh, Sim Kuan
Tang, Ho-Kin
He, Daojing
Zhang, Min
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
While fine-tuning pretrained models has become common practice, these models often underperform outside their specific domains. Recently developed model merging techniques enable the direct integration of multiple models, each fine-tuned for distinct tasks, into a single model. This strategy promotes multitasking capabilities without requiring retraining on the original datasets. However, existing methods fall short in addressing potential conflicts and complex correlations between tasks, especially in parameter-level adjustments, posing a challenge in effectively balancing parameter competition across various tasks. This paper introduces an innovative technique named PCB-Merging (Parameter Competition Balancing), a lightweight and training-free technique that adjusts the coefficients of each parameter for effective model merging. PCB-Merging employs intra-balancing to gauge parameter significance within individual tasks and inter-balancing to assess parameter similarities across different tasks. Parameters with low importance scores are dropped, and the remaining ones are rescaled to form the final merged model. We assessed our approach in diverse merging scenarios, including cross-task, cross-domain, and cross-training configurations, as well as out-of-domain generalization. The experimental results reveal that our approach achieves substantial performance enhancements across multiple modalities, domains, model sizes, number of tasks, fine-tuning forms, and large language models, outperforming existing model merging methods. The code is publicly available at: \url{https://github.com/duguodong7/pcb-merging}.
title Parameter Competition Balancing for Model Merging
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2410.02396