Machine Translation On Wmt2016 Romanian
Métriques
BLEU score
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | BLEU score |
|---|---|
| incorporating-a-local-translation-mechanism | 31.24 |
| finetuned-language-models-are-zero-shot | 38.1 |
| alleviating-the-inequality-of-attention-heads | 32.85 |
| edinburgh-neural-machine-translation-systems | 33.3 |
| finetuned-language-models-are-zero-shot | 37.3 |
| non-autoregressive-neural-machine-translation-1 | 31.44 |
| textbox-2-0-a-text-generation-library-with | - |
| cross-lingual-language-model-pretraining | 35.3 |
| flowseq-non-autoregressive-conditional | 32.03 |
| flowseq-non-autoregressive-conditional | 32.91 |
| deterministic-non-autoregressive-neural | 30.30 |
| flowseq-non-autoregressive-conditional | 32.46 |
| language-models-not-just-for-pre-training | 40.3 |
| gentranslate-large-language-models-are | 33.5 |
| flowseq-non-autoregressive-conditional | 30.69 |
| flowseq-non-autoregressive-conditional | 30.16 |
| adaptively-sparse-transformers | 33.1 |
| alleviating-the-inequality-of-attention-heads | 32.95 |
| levenshtein-transformer | 33.26 |
| incorporating-a-local-translation-mechanism | 33.26 |
| adaptively-sparse-transformers | 32.89 |