ShadowPEFT: Shadow Network for Parameter-Efficient Fine-Tuning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Xianming, Li, Zongxi, Lee, Tsz-fung Andrew, Li, Jing, Xie, Haoran, Li, Qing
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908983338467328
author Li, Xianming
Li, Zongxi
Lee, Tsz-fung Andrew
Li, Jing
Xie, Haoran
Li, Qing
author_facet Li, Xianming
Li, Zongxi
Lee, Tsz-fung Andrew
Li, Jing
Xie, Haoran
Li, Qing
contents Parameter-efficient fine-tuning (PEFT) reduces the training cost of full-parameter fine-tuning for large language models (LLMs) by training only a small set of task-specific parameters while freezing the pretrained backbone. However, existing approaches, such as Low-Rank Adaptation (LoRA), achieve adaptation by inserting independent low-rank perturbations directly to individual weights, resulting in a local parameterization of adaptation. We propose ShadowPEFT, a centralized PEFT framework that instead performs layer-level refinement through a depth-shared shadow module. At each transformer layer, ShadowPEFT maintains a parallel shadow state and evolves it repeatedly for progressively richer hidden states. This design shifts adaptation from distributed weight-space perturbations to a shared layer-space refinement process. Since the shadow module is decoupled from the backbone, it can be reused across depth, independently pretrained, and optionally deployed in a detached mode, benefiting edge computing scenarios. Experiments on generation and understanding benchmarks show that ShadowPEFT matches or outperforms LoRA and DoRA under comparable trainable-parameter budgets. Additional analyses on shadow pretraining, cross-dataset transfer, parameter scaling, inference latency, and system-level evaluation suggest that centralized layer-space adaptation is a competitive and flexible alternative to conventional low-rank PEFT.
format Preprint
id arxiv_https___arxiv_org_abs_2604_19254
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ShadowPEFT: Shadow Network for Parameter-Efficient Fine-Tuning
Li, Xianming
Li, Zongxi
Lee, Tsz-fung Andrew
Li, Jing
Xie, Haoran
Li, Qing
Computation and Language
Artificial Intelligence
Parameter-efficient fine-tuning (PEFT) reduces the training cost of full-parameter fine-tuning for large language models (LLMs) by training only a small set of task-specific parameters while freezing the pretrained backbone. However, existing approaches, such as Low-Rank Adaptation (LoRA), achieve adaptation by inserting independent low-rank perturbations directly to individual weights, resulting in a local parameterization of adaptation. We propose ShadowPEFT, a centralized PEFT framework that instead performs layer-level refinement through a depth-shared shadow module. At each transformer layer, ShadowPEFT maintains a parallel shadow state and evolves it repeatedly for progressively richer hidden states. This design shifts adaptation from distributed weight-space perturbations to a shared layer-space refinement process. Since the shadow module is decoupled from the backbone, it can be reused across depth, independently pretrained, and optionally deployed in a detached mode, benefiting edge computing scenarios. Experiments on generation and understanding benchmarks show that ShadowPEFT matches or outperforms LoRA and DoRA under comparable trainable-parameter budgets. Additional analyses on shadow pretraining, cross-dataset transfer, parameter scaling, inference latency, and system-level evaluation suggest that centralized layer-space adaptation is a competitive and flexible alternative to conventional low-rank PEFT.
title ShadowPEFT: Shadow Network for Parameter-Efficient Fine-Tuning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2604.19254