PyTorch implementation of Few-Shot Domain Adaptation (https://arxiv.org/abs/1711.02536)
Disclaimer: I'm not in any way affiliated with the authors. There might be errors in the implementation.
- Main algorithm implemented
- Fix TODOs in code
- Comprehensive tests
- More datasets to test on (currently only MNIST -> USPS)
- The authors don't give the value of γ they used, one should be found via cross-validation
Use python3 main.py to run the MNIST -> SVHN training and print accuracy at each epoch.
Preliminary results show ~46% accuracy on test set with n=7 samples per class from target domain. This approximately matches what is reported in the paper (47.0%)