Unsupervised Machine Translation On Wmt2016
Métriques
BLEU
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | BLEU |
|---|---|
| unsupervised-neural-machine-translation-with | 21.7 |
| mass-masked-sequence-to-sequence-pre-training | 28.3 |
| language-models-are-few-shot-learners | 29.7 |
| cross-lingual-language-model-pretraining | 26.4 |
| phrase-based-neural-unsupervised-machine | 20.2 |
| unsupervised-neural-machine-translation-1 | 20.0 |
| an-effective-approach-to-unsupervised-machine | 26.9 |