Includes additional materials for the following keras.io blog post: Semi-supervision and domain adaptation with AdaMatch.
AdaMatch.ipynb
: Original notebook submitted for the PR.Vanilla_WideResNet.ipynb
: Trains a WideResNet-28-2 on MNIST (source domain) and evaluates on the SVHN dataset (target domain). The model trained in this notebook serves as the baseline.
- François Chollet for helping with the implementation.
- ML-GDE program for providing GCP credits that supported the experiments.
@misc{berthelot2021adamatch,
title={AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptation},
author={David Berthelot and Rebecca Roelofs and Kihyuk Sohn and Nicholas Carlini and Alex Kurakin},
year={2021},
eprint={2106.04732},
archivePrefix={arXiv},
primaryClass={cs.LG}
}