Official implementation for ATDOC

[CVPR-2021] Domain Adaptation with Auxiliary Target Domain-Oriented Classifier

[Update @ Nov 23 2021]

1. [For Office, please change the max-epoch to 100; for VISDA-C, change the max-epoch to 1 and change the net to resnet101]
2. Add the code associated with SSDA, change the max-epoch to 20 for DomainNet-126
3. Thank @lyxok1 for pointing out the typo in Eq.(6), we have corrected it in the new verison of this paper.

Below is the demo for ATDOC on a UDA task of Office-Home [max_epoch to 50]:

1. installing packages

   python == 3.6.8 pytorch ==1.1.0 torchvision == 0.3.0 numpy, scipy, sklearn, PIL, argparse, tqdm

2. download the Office-Home dataset

   mkdir dataset

   cd dataset

   pip install gdown

   gdown https://drive.google.com/u/0/uc?id=0B81rNlvomiwed0V1YUxQdC1uOTg&export=download

   unzip OfficeHomeDataset_10072016.zip

   mv ./OfficeHomeDataset_10072016/Real\ World ./OfficeHomeDataset_10072016/RealWorld

   cd ../

3. run the main file with 'Source-model-only'

   python demo_uda.py --pl none --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --method srconly --output logs/uda/run1/

4. run the main file with 'ATDOC-NC'

   python demo_uda.py --pl atdoc_nc --tar_par 0.1 --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --method srconly --output logs/uda/run1/

5. run the main file with 'ATDOC-NA'

   python demo_uda.py --pl atdoc_na --tar_par 0.2 --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --method srconly --output logs/uda/run1/

6. run the main file with 'ATDOC-NA' combined with 'CDAN+E'

   python demo_uda.py --pl atdoc_na --tar_par 0.2 --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --method CDANE --output logs/uda/run1/

7. run the main file with 'ATDOC-NA' combined with 'MixMatch'

   python demo_mixmatch.py --pl none --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --output logs/uda/run1/

8. run the main file with 'ATDOC-NA' combined with 'MixMatch'

   python demo_mixmatch.py --pl atdoc_na --dset office-home --max_epoch 50 --s 0 --t 1 --gpu_id 0 --output logs/uda/run1/

Citation

If you find this code useful for your research, please cite our paper

@inproceedings{liang2021domain,
    title={Domain Adaptation with Auxiliary Target Domain-Oriented Classifier},
    author={Liang, Jian and Hu, Dapeng and Feng, Jiashi},
    booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
    year={2021}
}

Contact

• liangjian92@gmail.com
• dapeng.hu@u.nus.edu
• elefjia@nus.edu.sg