Optimal transport implementation for domain adaptaion for cross domain reccomendation
- Jupyter Notebook 50.9%
- Python 48.9%
- Shell 0.2%
| .vscode | ||
| DArec_opt | ||
| data | ||
| examples | ||
| experiment_framework | ||
| I-DArec | ||
| reference | ||
| U-DArec | ||
| .gitignore | ||
| .gitmodules | ||
| .python-version | ||
| pyproject.toml | ||
| README.md | ||
| run_experiments.py | ||
| uv.lock | ||
| visualize_embeddings.ipynb | ||
Running this yourself
uv sync
you want to start by downloading the data. No Guarentees that the links still work:
./data/download.sh
Then you want to run the preprocessing to generate the correct numpy matricies. You can run the original DARec code if you would like using the corresponding Data_Preprocessing.py or you can trust I implemented the multi-mode version and it all works with:
python DArec_opt/Data_Preprocessing.py