This file introduce the usage and outline of code.
The project code has the following directory structure
code
│ ADDA.py # implementation of ADDA
│ Adversarial.py
│ DANN.py # implementation of DANN
│ Dataset.py # implementation of Dataset
│ draw.py # draw plot function
│ implementation_PR_PL.py # implementation of PR-PL
│ main.py
│ models.py # implementation of models
│ model_PR_PL.py # implementation of PR-PL model
│ parser.py
│ README.md # this file
│ ResNet.py # implementation of ResNet model
│ search.py # implementation of Bayes Search
│ SVM.py # implementation of SVM
│ tca.py # implementation of TCA
│
├─SEED-IV # Train Data
To run the baseline and our model, go to path /code, run command:
python main.py --model XXX
the model arg have the following selection:
- Conventional ML model:
- SVM
- Conventional DL models:
- MLP
- resnet
- Domain generalization models:
- IRM
- Domain adaptation models:
- tca
- DANN
- ADDA
- prpl
You can also use args like --lr
or --batch_size
to tune the hyperparameters.
All experiments are done on i9-10900X CPU @ 3.70GHz along with a GeForce RTX 3090 GPU.