Adversarial Partial Domain Adaptation by Cycle Inconsistency

PyTorch implementation of the ECCV2022 paper “Adversarial Partial Domain Adaptation by Cycle Inconsistency” [paper]

Environments

• Python: 3.7.3
• PyTorch: 1.8.1

Training

Data preparation

By running the script examples/domain_adaptation/partial/run.sh, datasets will be downloaded automatically to the directory examples/domain_adaptation/partial/data/.

Or you can download the datasets and put them into the right directory.

Run training scripts

cd examples/domain_adaptation/partial/
bash run.sh

Acknowledgement

• This repository is heavily based on the codebase thuml/Transfer-Learning-Library.

• If you find this paper/code useful, please consider citing us:

@inproceedings{lin2022partial,
  author       = {Kun-Yu Lin and Jiaming Zhou and Yukun Qiu and Wei-Shi Zheng},
  title        = {Adversarial Partial Domain Adaptation by Cycle Inconsistency},
  booktitle    = {European Conference on Computer Vision},
  year         = {2022},
}