Machine Translation On Wmt2016 German English
Metrics
BLEU score
Results
Performance results of various models on this benchmark
| Paper Title | ||
|---|---|---|
| FLAN 137B (few-shot, k=11) | 40.7 | Finetuned Language Models Are Zero-Shot Learners |
| FLAN 137B (zero-shot) | 38.9 | Finetuned Language Models Are Zero-Shot Learners |
| Attentional encoder-decoder + BPE | 38.6 | Edinburgh Neural Machine Translation Systems for WMT 16 |
| Linguistic Input Features | 32.9 | Linguistic Input Features Improve Neural Machine Translation |
| SMT + iterative backtranslation (unsupervised) | 23.05 | Unsupervised Statistical Machine Translation |
| Unsupervised NMT + weight-sharing | 14.62 | Unsupervised Neural Machine Translation with Weight Sharing |
| Unsupervised S2S with attention | 13.33 | Unsupervised Machine Translation Using Monolingual Corpora Only |
| Exploiting Mono at Scale (single) | - | Exploiting Monolingual Data at Scale for Neural Machine Translation |
0 of 8 row(s) selected.