Tokenization

Subword Regularization

A data augmentation technique applying different possible segmentations of the same text during training, improving model robustness and generalization.

← Zurück