Unsupervised Machine Translation On Wmt2016 2
Metrics
BLEU
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | BLEU |
|---|---|
| language-models-are-few-shot-learners | 21 |
| cross-lingual-language-model-pretraining | 33.3 |
| mass-masked-sequence-to-sequence-pre-training | 35.2 |