https://github.com/pris-cv/mutual-to-separate
Code release for Mind the Gap: Enlarging the Domain Gap in Open Set Domain Adaptation (TCSVT 2023)
https://github.com/pris-cv/mutual-to-separate
domain-adaptation open-set open-set-domain-adaptation open-set-recognition
Last synced: 26 days ago
JSON representation
Code release for Mind the Gap: Enlarging the Domain Gap in Open Set Domain Adaptation (TCSVT 2023)
- Host: GitHub
- URL: https://github.com/pris-cv/mutual-to-separate
- Owner: PRIS-CV
- License: mit
- Created: 2020-03-08T06:44:56.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2023-10-20T14:04:14.000Z (over 1 year ago)
- Last Synced: 2025-04-24T05:46:12.279Z (26 days ago)
- Topics: domain-adaptation, open-set, open-set-domain-adaptation, open-set-recognition
- Language: Python
- Homepage:
- Size: 55.7 KB
- Stars: 22
- Watchers: 6
- Forks: 4
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework
Code release for Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework [TCSVT 2023]
## Dataset
### OFFICE-Home
## Requirements
- python 3.6
- PyTorch 1.1.0
- torchvision 0.3.0
- Tensorflow 1.9.0
- Tensorlayer 1.11
- Tensorboard
- tensorpack
## GPU Version
- 1080ti
## Training
- Download datasets
- Train: `python Office_Home.py "Art" "Clipart" "0" "A_C" 0.2 0.2`
- Description: PyTorch Open-set OFFICE-HOME Training with ResNet50 (PRE-TRAINED WITH IMAGENET).
## Reference codes
**https://github.com/thuml/easydl**
## Citation
If you find this paper useful in your research, please consider citing:
```
@ARTICLE{chang_Mind,
author={Dongliang Chang, Aneeshan Sain, Zhanyu Ma, Yi-Zhe Song, Ruiping Wang, and Jun Guo},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
title={Mind the Gap: Open Set Domain Adaptation via Mutual-to-Separate Framework},
year={2023},
doi={10.1109/TCSVT.2023.3326862}}
```
## Contact
- [email protected]
- [email protected]