Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2402.08249 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911878955925504 |
|---|---|
| author | Jin, Ying Wang, Jiaqi Lin, Dahua |
| author_facet | Jin, Ying Wang, Jiaqi Lin, Dahua |
| contents | We consider multi-source free domain adaptation, the problem of adapting multiple existing models to a new domain without accessing the source data. Among existing approaches, methods based on model ensemble are effective in both the source and target domains, but incur significantly increased computational costs. Towards this dilemma, in this work, we propose a novel framework called SepRep-Net, which tackles multi-source free domain adaptation via model Separation and Reparameterization.Concretely, SepRep-Net reassembled multiple existing models to a unified network, while maintaining separate pathways (Separation). During training, separate pathways are optimized in parallel with the information exchange regularly performed via an additional feature merging unit. With our specific design, these pathways can be further reparameterized into a single one to facilitate inference (Reparameterization). SepRep-Net is characterized by 1) effectiveness: competitive performance on the target domain, 2) efficiency: low computational costs, and 3) generalizability: maintaining more source knowledge than existing solutions. As a general approach, SepRep-Net can be seamlessly plugged into various methods. Extensive experiments validate the performance of SepRep-Net on mainstream benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_08249 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SepRep-Net: Multi-source Free Domain Adaptation via Model Separation And Reparameterization Jin, Ying Wang, Jiaqi Lin, Dahua Computer Vision and Pattern Recognition We consider multi-source free domain adaptation, the problem of adapting multiple existing models to a new domain without accessing the source data. Among existing approaches, methods based on model ensemble are effective in both the source and target domains, but incur significantly increased computational costs. Towards this dilemma, in this work, we propose a novel framework called SepRep-Net, which tackles multi-source free domain adaptation via model Separation and Reparameterization.Concretely, SepRep-Net reassembled multiple existing models to a unified network, while maintaining separate pathways (Separation). During training, separate pathways are optimized in parallel with the information exchange regularly performed via an additional feature merging unit. With our specific design, these pathways can be further reparameterized into a single one to facilitate inference (Reparameterization). SepRep-Net is characterized by 1) effectiveness: competitive performance on the target domain, 2) efficiency: low computational costs, and 3) generalizability: maintaining more source knowledge than existing solutions. As a general approach, SepRep-Net can be seamlessly plugged into various methods. Extensive experiments validate the performance of SepRep-Net on mainstream benchmarks. |
| title | SepRep-Net: Multi-source Free Domain Adaptation via Model Separation And Reparameterization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2402.08249 |