Source code for my Thesis - 2023.2 - Building an Efficient Retrieval System for Vietnamese Legal Domain. See my repo for the latest updated version!
Raw, data and trained models are upload on OneDrive
T ran the code in Kaggle, I recommend you to upload the notebooks in \kaggle_notebooks
folder and relevant dataset and trained models to Kaggle to run. You need to create a Wandb account and get the key to run the RetroMAE pretraining and Cross-encoder training notebooks.