Helsinki-NLP/opus-mt-en-vi

• This model is a fine-tune checkpoint of Helsinki-NLP/opus-mt-en-vi.
• This model reaches BLEU score = 33.086 on the test set of IWSLT'15 English-Vietnamese data.

Fine-tuning hyper-parameters

• learning_rate = 1e-4
• batch_size = 4
• num_train_epochs = 3.0