Mada-SGD

A novel optimization-based meta learning algorithm for few-shot SAR target recognition.

The code in this toolbox implements the "Few-Shot SAR Target Recognition Through Meta-Adaptive Hyperparameters Learning for Fast Adaptation" in IEEE Transactions on Geoscience and Remote Sensing (TGRS). More specifically, it is detailed as follow.

requirements

---python 3.7
---pytorch >= 1.9
---CUDA 10.2

trian

•       trian.py
  

test

•       test.py
  

Citation

please kindly cite this paper if our Mada-SGD can give you any inspiration for your research, thanks a lot.

Z. Zeng, J. Sun, Y. Wang, D. Gu, Z. Han and W. Hong, "Few-Shot SAR Target Recognition Through Meta-Adaptive Hyperparameters Learning for Fast Adaptation," in IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1-17, 2023, Art no. 5219517, doi: 10.1109/TGRS.2023.3325988.

Contact

Zhiqiang Zeng
Email:[email protected]

references

1. the original MSTAR dataset information: https://www.sdms.afrl.af.mil/
2. https://github.com/dragen1860/Reptile-Pytorch
3. https://github.com/RL-VIG/LibFewShot
4. https://github.com/learnables/learn2learn
5. https://github.com/dragen1860/MAML-Pytorch