Machine Translation on IT En-De out-of-domain WMT14 (test)
27.5BLEUAT
Evaluation results
| Method | Links | |
|---|---|---|
| ATModel Type=Autoregressive Teacher2021.06 | 27.5 | |
| Levenshtein Transformer + LFRModel Type=Non-Autoregressive Transformer, Strategy=Low-Frequency Rejuvenation2021.06 | 25.9 | |
| Mask-Predict + LFRModel Type=Non-Autoregressive Transformer, Strategy=Low-Frequency Rejuvenation2021.06 | 25.4 | |
| Levenshtein TransformerModel Type=Non-Autoregressive Transformer2021.06 | 24.7 | |
| Mask-PredictModel Type=Non-Autoregressive Transformer2021.06 | 24.6 | |
| Byte-nCFModel Type=DENSE, Adaptation Mode=Zero-shot cross-domain adaptation2022.05 | 11.6 | |
| Byte-nCFModel Type=ONE-HOT, Adaptation Mode=Zero-shot cross-domain adaptation2022.05 | 11.3 | |
| ByteModel Type=DENSE, Adaptation Mode=Zero-shot cross-domain adaptation2022.05 | 10.4 | |
| ByteModel Type=ONE-HOT, Adaptation Mode=Zero-shot cross-domain adaptation2022.05 | 10.2 | |
| SubwordModel Type=DENSE, Adaptation Mode=Zero-shot cross-domain adaptation2022.05 | 3.9 |