This is an UNOFFICIAL implemnetation of Co-regularized Alignment for Unsupervised Domain Adaptation
- python 3.6
- PyTorch 1.1
- tensorboardX
- numpy
- cv2
- tqdm
agreement loss | diverse loss |
---|---|
net_num | validation | test |
---|---|---|
net1 | ||
net2 |
TODO:
- DIRT-T model
Note: See https://github.com/lyakaap/VAT-pytorch for original implementation of VAT loss