Split-Attention-dataset

The dataset includes the raw dataset, preprocessed data that is suitable for model training (data-bin), the best-trained model (checkpoints, including training and inference logs), and preprocessed syntactic dependency triplets (dep-relation).