| --- |
| license: apache-2.0 |
| tags: |
| - remote-sensing |
| - long-tailed |
| - parameter-efficient-fine-tuning |
| - peft |
| - lora |
| - meta-learning |
| - cvpr2025 |
| - arxiv:2603.01759 |
| --- |
| |
| # MetaPEFT checkpoints (CVPR 2025 Highlight) |
|
|
| Model weights for the CVPR 2025 **Highlight** paper |
| **"Meta-Learning Hyperparameters for Parameter Efficient Fine-Tuning"** |
| by Zichen Tian, Yaoyao Liu and Qianru Sun. |
|
|
| <!-- TODO Zichen: 一句话描述这个 repo 里是什么 checkpoint（backbone、在哪个 split 上训的），我不知道就没写 --> |
|
|
| ## Paper |
|
|
| - **Title:** Meta-Learning Hyperparameters for Parameter Efficient Fine-Tuning |
| - **Venue:** CVPR 2025 (Highlight), pp. 23037-23047 |
| - **arXiv:** https://arxiv.org/abs/2603.01759 |
| - **CVF Open Access:** https://openaccess.thecvf.com/content/CVPR2025/html/Tian_Meta-Learning_Hyperparameters_for_Parameter_Efficient_Fine-Tuning_CVPR_2025_paper.html |
| - **Project page:** https://www.doem1997.com/metapeft/ |
| - **Code:** https://github.com/doem97/metalora |
| - **Dataset:** https://huggingface.co/datasets/doem1997/rs_lt |
|
|
| ## Abstract |
|
|
| Training large foundation models from scratch for domain-specific applications is almost impossible due to data limits and long-tailed distributions -- taking remote sensing (RS) as an example. Fine-tuning natural image pre-trained models on RS images is a straightforward solution. To reduce computational costs and improve performance on tail classes, existing methods apply parameter-efficient fine-tuning (PEFT) techniques, such as LoRA and AdaptFormer. However, we observe that fixed hyperparameters -- such as intra-layer positions, layer depth, and scaling factors, can considerably hinder PEFT performance, as fine-tuning on RS images proves highly sensitive to these settings. To address this, we propose MetaPEFT, a method incorporating adaptive scalers that dynamically adjust module influence during fine-tuning. MetaPEFT dynamically adjusts three key factors of PEFT on RS images: module insertion, layer selection, and module-wise learning rates, which collectively control the influence of PEFT modules across the network. We conduct extensive experiments on three transfer-learning scenarios and five datasets in both RS and natural image domains. The results show that MetaPEFT achieves state-of-the-art performance in cross-spectral adaptation, requiring only a small amount of trainable parameters and improving tail-class accuracy significantly. |
|
|
| ## Citation |
|
|
| ```bibtex |
| @InProceedings{Tian_2025_CVPR, |
| author = {Tian, Zichen and Liu, Yaoyao and Sun, Qianru}, |
| title = {Meta-Learning Hyperparameters for Parameter Efficient Fine-Tuning}, |
| booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, |
| month = {June}, |
| year = {2025}, |
| pages = {23037-23047} |
| } |
| ``` |
|
|
