TSFLora: Token-Compressed Split Fine-Tuning for Wireless Edge Networks
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914593708703744 |
|---|---|
| author | Qiang, Xianke Chang, Zheng Wang, Li Liang, Ying-Chang |
| author_facet | Qiang, Xianke Chang, Zheng Wang, Li Liang, Ying-Chang |
| contents | Adapting large AI models (LAMs) to personalized edge data is challenging because wireless devices have limited memory, computation, and uplink capacity. Federated fine-tuning preserves data privacy but still requires each device to host the full model, while split learning reduces device memory at the cost of heavy activation transmission. This paper proposes TSFLora, a token-compressed split fine-tuning framework for communication-efficient LAM adaptation at the edge. TSFLora combines attention-guided token selection, token merging, low-bit activation quantization, and LoRA-based adaptation within a split federated training pipeline. The key idea is to compress the intermediate token sequence before transmission so that the system reduces both uplink traffic and server-side processing without changing the frozen backbone. Experiments on ViT models over CIFAR-10, CIFAR-100, and TinyImageNet show that TSFLora achieves up to \textbf{6.8$\times$} communication reduction and \textbf{41\%} memory saving while maintaining competitive accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_23988 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TSFLora: Token-Compressed Split Fine-Tuning for Wireless Edge Networks Qiang, Xianke Chang, Zheng Wang, Li Liang, Ying-Chang Distributed, Parallel, and Cluster Computing Machine Learning Adapting large AI models (LAMs) to personalized edge data is challenging because wireless devices have limited memory, computation, and uplink capacity. Federated fine-tuning preserves data privacy but still requires each device to host the full model, while split learning reduces device memory at the cost of heavy activation transmission. This paper proposes TSFLora, a token-compressed split fine-tuning framework for communication-efficient LAM adaptation at the edge. TSFLora combines attention-guided token selection, token merging, low-bit activation quantization, and LoRA-based adaptation within a split federated training pipeline. The key idea is to compress the intermediate token sequence before transmission so that the system reduces both uplink traffic and server-side processing without changing the frozen backbone. Experiments on ViT models over CIFAR-10, CIFAR-100, and TinyImageNet show that TSFLora achieves up to \textbf{6.8$\times$} communication reduction and \textbf{41\%} memory saving while maintaining competitive accuracy. |
| title | TSFLora: Token-Compressed Split Fine-Tuning for Wireless Edge Networks |
| topic | Distributed, Parallel, and Cluster Computing Machine Learning |
| url | https://arxiv.org/abs/2605.23988 |