paper-with-me

Papers

Task Singular Vectors: Reducing Task Interference in Model Merging

2024-11-26 · CVPR 2025 1 · Antonio Andrea Gargiulo, Donato Crisostomi, Maria Sofia Bucarelli, Simone Scardapane, Fabrizio Silvestri, Emanuele Rodolà

Task Arithmetic has emerged as a simple yet effective method to merge models without additional training. However, by treating entire networks as flat parameter vectors, it overlooks key structural information and is susceptible to task interference. In this paper, we study task vectors at the layer level, focusing on task layer matrices and their singular value decomposition. In particular, we concentrate on the resulting singular vectors, which we refer to as Task Singular Vectors (TSV). Recognizing that layer task matrices are often low-rank, we propose TSV-Compress (TSV-C), a simple procedure that compresses them to 10% of their original size while retaining 99% of accuracy. We further leverage this low-rank space to define a new measure of task interference based on the interaction of singular vectors from different tasks. Building on these findings, we introduce TSV-Merge (TSV-M), a novel model merging approach that combines compression with interference reduction, significantly outperforming existing methods.

📄 PDF Abstract BibTeX arXiv:2412.00081

Code (1)

AntoAndGar/task_singular_vectors 공식 구현 pytorch

Tasks

ClassificationImage ClassificationMulti-Label Image RecognitionTask Arithmetic

Similar Papers 제목 키워드 기반

Beyond Traditional Beamforming: Singular Vector Projection for MU-MIMO

2023-11-07 · Md Saheed Ullah, Rafid Umayer Murshed, Mohammad Saquib, Md. Forkan Uddin

This letter introduces a low-complexity beamforming approach for MU-MIMO systems with multiple data streams per user, minimizing inter-user interference and improving spectral efficiency (SE). The Interference-Optimized …

Computational EfficiencyManagement

AdaRank: Adaptive Rank Pruning for Enhanced Model Merging

2025-03-28 · Chanhyuk Lee, Jiho Choi, Chanryeol Lee, Donggyun Kim 외

Model merging has emerged as a promising approach for unifying independently fine-tuned models into an integrated framework, significantly enhancing computational efficiency in multi-task learning. Recently, several SVD-…

Computational EfficiencymodelMulti-Task Learning

Revisiting Weight Averaging for Model Merging

2024-12-11 · Jiho Choi, Donggyun Kim, Chanhyuk Lee, Seunghoon Hong

Model merging aims to build a multi-task learner by combining the parameters of individually fine-tuned models without additional training. While a straightforward approach is to average model parameters across tasks, th…

image-classificationImage ClassificationmodelMulti-Task Learning

LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging

2025-02-15 · Zehua Liu, Han Wu, Yuxuan Yao, Ruifeng She 외

While most current approaches rely on further training techniques, such as fine-tuning or reinforcement learning, to enhance model capacities, model merging stands out for its ability of improving models without requirin…

Language ModelingLanguage ModellingLarge Language Model

MultiLoRA: Democratizing LoRA for Better Multi-Task Learning

2023-11-20 · Yiming Wang, Yu Lin, Xiaodong Zeng, Guannan Zhang

LoRA achieves remarkable resource efficiency and comparable performance when adapting LLMs for specific tasks. Since ChatGPT demonstrated superior performance on various tasks, there has been a growing desire to adapt on…

Multi-Task LearningNatural Language UnderstandingWorld Knowledge