ATM: Improving Model Merging by Alternating Tuning and Merging
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911096942624768 |
|---|---|
| author | Zhou, Luca Solombrino, Daniele Crisostomi, Donato Bucarelli, Maria Sofia Silvestri, Fabrizio Rodolà, Emanuele |
| author_facet | Zhou, Luca Solombrino, Daniele Crisostomi, Donato Bucarelli, Maria Sofia Silvestri, Fabrizio Rodolà, Emanuele |
| contents | Model merging has emerged as a cost-efficient approximation to multitask learning. Among merging strategies, task arithmetic is notable for its simplicity and effectiveness. In this work, we provide a theoretical motivation for task vectors by highlighting that, under single-epoch full-batch gradient descent, they are equivalent to multitask gradients. This insight leads us to reinterpret model merging as a single step in an iterative procedure that Alternates between Tuning and Merging (ATM). We propose two applications of ATM: (1) as an alternative to multitask learning in scenarios where data sharing is restricted (e.g., federated settings), and (2) as a lightweight refinement step to improve existing model merging methods using a small validation set. Experiments across diverse vision tasks demonstrate the effectiveness of ATM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_03055 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | ATM: Improving Model Merging by Alternating Tuning and Merging Zhou, Luca Solombrino, Daniele Crisostomi, Donato Bucarelli, Maria Sofia Silvestri, Fabrizio Rodolà, Emanuele Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition Model merging has emerged as a cost-efficient approximation to multitask learning. Among merging strategies, task arithmetic is notable for its simplicity and effectiveness. In this work, we provide a theoretical motivation for task vectors by highlighting that, under single-epoch full-batch gradient descent, they are equivalent to multitask gradients. This insight leads us to reinterpret model merging as a single step in an iterative procedure that Alternates between Tuning and Merging (ATM). We propose two applications of ATM: (1) as an alternative to multitask learning in scenarios where data sharing is restricted (e.g., federated settings), and (2) as a lightweight refinement step to improve existing model merging methods using a small validation set. Experiments across diverse vision tasks demonstrate the effectiveness of ATM. |
| title | ATM: Improving Model Merging by Alternating Tuning and Merging |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2411.03055 |