Machine Translation on WMT17 Zh-En (test)
26.5BLEUStrokeNet
Evaluation Results
| Method | Links | |
|---|---|---|
| StrokeNet2022.11 | 26.5 | |
| Prior2022.11 | 25.5 | |
| Norm2022.11 | 25.3 | |
| Vanilla2022.11 | 24.4 | |
| Pretrained Transformer LargeInitialization=N/A2021.03 | 24.2 | |
| ELUInitialization=Pretrain, Feature Size k (cross)=64, Feature Size k (causal)=64, Train Time (GPU hours)=80h2021.03 | 23.8 | |
| T2RInitialization=Pretrain, Feature Size k (cross)=32, Feature Size k (causal)=4, Train Time (GPU hours)=82h2021.03 | 23.8 | |
| ELUInitialization=Random Init., Feature Size k (cross)=64, Feature Size k (causal)=64, Train Time (GPU hours)=120h2021.03 | 23.4 | |
| RFAInitialization=Random Init., Feature Size k (cross)=32, Feature Size k (causal)=4, Train Time (GPU hours)=135h2021.03 | 23.4 | |
| RFAInitialization=Pretrain, Feature Size k (cross)=32, Feature Size k (causal)=4, Train Time (GPU hours)=90h2021.03 | 23.2 | |
| T2RInitialization=Random Init., Feature Size k (cross)=32, Feature Size k (causal)=4, Train Time (GPU hours)=123h2021.03 | 23.1 |