Fetching the paper…

Explaining Sequence-Level Knowledge Distillation as Data-Augmentation for Neural Machine Translation · Around