Machine Translation on TED (test)
62.7ChrF++BPE baseline
Evaluation Results
| Method | Links | |
|---|---|---|
| BPE baselineArchitecture=Softmax, L_base=5, d_model=3002024.08 | 62.7 | |
| Multilingual SoftmaxArchitecture=Softmax, Trained on=en, de, es, fr, it, L_base=5, d_model=3002024.08 | 62.6 | |
| Multilingual SoftmaxArchitecture=Softmax, Trained on=en, de, es, fr, it, ru, uk, pl, vi, zh, L_base=10, d_model=5762024.08 | 61.5 | |
| vMF-NMTArchitecture=vMF, Trained on=es, en, L_base=1, d_model=3002024.08 | 60.9 | |
| vMF-NMTArchitecture=vMF, Trained on=es+en, L_base=2, d_model=3002024.08 | 59.9 | |
| Supervised fine-tuneArchitecture=vMF, L_base=5, d_model=3002024.08 | 59.9 | |
| vMF-NMTArchitecture=vMF, Trained on=en, de, es, fr, it, L_base=5, d_model=3002024.08 | 59.5 | |
| vMF-NMTArchitecture=vMF, Trained on=en, de, es, fr, it, ru, uk, pl, L_base=8, d_model=3002024.08 | 57.8 | |
| vMF-NMTArchitecture=vMF, Trained on=en, de, es, fr, it, ru, uk, pl, vi, zh, L_base=10, d_model=3002024.08 | 57.2 | |
| vMF-NMTArchitecture=vMF, Trained on=en, de, fr, it, L_base=4, d_model=3002024.08 | 55.1 | |
| Softmax (Garcia et al., 2021)Architecture=Softmax, L_base=5, d_model=3002024.08 | 48.9 | |
| Supervised fine-tuneArchitecture=Softmax, L_base=5, d_model=3002024.08 | 46.9 |