Weight Weaving: Parameter Pooling for Data-Free Model Merging

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chaves, Levy, Valle, Eduardo, Avila, Sandra
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917018336231424
author Chaves, Levy
Valle, Eduardo
Avila, Sandra
author_facet Chaves, Levy
Valle, Eduardo
Avila, Sandra
contents Model merging provides a cost-effective and data-efficient combination of specialized deep neural networks through parameter integration. This technique leverages expert models across downstream tasks without requiring retraining. Most model merging approaches critically depend on scaling hyper-parameters $λ$, which weight each model's contribution globally or individually. Principled approaches for setting scaling factors without accessing any data (data-free) are scarce, often leading researchers to tune $λ$ using privileged data from the evaluation set, which is obviously unfeasible in practice. To address this limitation, we introduce Weight Weaving, a plug-and-play technique that pools model weights across $λ$ values search space using user-defined pooling functions, such as averaging, random selection, or even existing model merging methods. Our method demonstrates high modularity, imposing minimal constraints on the search space. It operates orthogonally to existing model merging methods and eliminates evaluation data requirements. We validate Weight Weaving across three ViT variants in three experimental setups: vision multi-task learning, vision continual learning, and domain generalization. Our method consistently improves the performance of several model merging methods, achieving average accuracy gains of up to 15.9 percentage points in a data-free setting.
format Preprint
id arxiv_https___arxiv_org_abs_2510_13921
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Weight Weaving: Parameter Pooling for Data-Free Model Merging
Chaves, Levy
Valle, Eduardo
Avila, Sandra
Machine Learning
Computer Vision and Pattern Recognition
Model merging provides a cost-effective and data-efficient combination of specialized deep neural networks through parameter integration. This technique leverages expert models across downstream tasks without requiring retraining. Most model merging approaches critically depend on scaling hyper-parameters $λ$, which weight each model's contribution globally or individually. Principled approaches for setting scaling factors without accessing any data (data-free) are scarce, often leading researchers to tune $λ$ using privileged data from the evaluation set, which is obviously unfeasible in practice. To address this limitation, we introduce Weight Weaving, a plug-and-play technique that pools model weights across $λ$ values search space using user-defined pooling functions, such as averaging, random selection, or even existing model merging methods. Our method demonstrates high modularity, imposing minimal constraints on the search space. It operates orthogonally to existing model merging methods and eliminates evaluation data requirements. We validate Weight Weaving across three ViT variants in three experimental setups: vision multi-task learning, vision continual learning, and domain generalization. Our method consistently improves the performance of several model merging methods, achieving average accuracy gains of up to 15.9 percentage points in a data-free setting.
title Weight Weaving: Parameter Pooling for Data-Free Model Merging
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.13921