Machine Translation on WMT 22 (Zh -> En) (test)
33.5BLEUWMT22 Winners
Evaluation results
| Method | Links | ||
|---|---|---|---|
| WMT22 Winners2024.06 | 33.5 | 81 | |
| Target-constrained tuning LoRABase Model=LLaMA2-7b-Chat, tuning_method=LoRA2024.06 | 25.3 | 78.5 | |
| TASTEBackbone=LLaMA-2-7b, Training Strategy=Tuning with Fixed Embedding Layer, Subtask Form=Text Classification2024.06 | 24.87 | 79.53 | |
| TIMBackbone=BLOOMZ-7b1-mt2024.06 | 24.51 | 79.71 | |
| TASTEBackbone=LLaMA-2-7b, Training Strategy=Tuning with Fixed Embedding Layer, Subtask Form=Quality Estimation2024.06 | 24.47 | 79.35 | |
| TASTEBackbone=BLOOMZ-7b1-mt, Training Strategy=Tuning with Fixed Embedding Layer, Subtask Form=Quality Estimation2024.06 | 24.41 | 80.4 | |
| Vanilla Instruction tuning LoRABase Model=LLaMA2-7b-Chat, tuning_method=LoRA2024.06 | 24.4 | 77.6 | |
| MT-FixEmbBackbone=LLaMA-2-7b, Training Strategy=Tuning with Fixed Embedding Layer2024.06 | 24.3 | 79.02 | |
| TASTEBackbone=LLaMA-2-7b, Training Strategy=Full-Parameter Tuning, Subtask Form=Quality Estimation2024.06 | 24.27 | 79.17 | |
| TASTEBackbone=LLaMA-2-7b, Training Strategy=Full-Parameter Tuning, Subtask Form=Text Classification2024.06 | 24.23 | 79.31 | |
| TASTEBackbone=BLOOMZ-7b1-mt, Training Strategy=Tuning with Fixed Embedding Layer, Subtask Form=Text Classification2024.06 | 24.2 | 80.28 | |
| MT-FullBackbone=LLaMA-2-7b, Training Strategy=Full-Parameter Tuning2024.06 | 23.8 | 78.72 | |
| MT-FixEmbBackbone=BLOOMZ-7b1-mt, Training Strategy=Tuning with Fixed Embedding Layer2024.06 | 23.43 | 79.84 | |
| TASTEBackbone=BLOOMZ-7b1-mt, Training Strategy=Full-Parameter Tuning, Subtask Form=Quality Estimation2024.06 | 23.15 | 79.36 | |
| TASTEBackbone=BLOOMZ-7b1-mt, Training Strategy=Full-Parameter Tuning, Subtask Form=Text Classification2024.06 | 23.04 | 79.14 | |
| MT-FullBackbone=BLOOMZ-7b1-mt, Training Strategy=Full-Parameter Tuning2024.06 | 22.81 | 79.25 | |
| ParroTBackbone=BLOOMZ-7b1-mt2024.06 | 22.7 | 79 | |
| NLLB-3.3b2024.06 | 21.07 | 76.92 | |
| BaylingBackbone=LLaMA-2-7b2024.06 | 20.31 | 77.48 | |
| ParroTBackbone=LLaMA-2-7b2024.06 | 20.2 | 75.9 | |
| Reweight AttentionBase Model=LLaMA2-7b-Chat2024.06 | 20.1 | 77.4 | |
| Vanilla ZeroshotBase Model=LLaMA2-7b-Chat2024.06 | 19.7 | 77.4 | |
| Contrastive DecodingBase Model=LLaMA2-7b-Chat2024.06 | 19.7 | 77.5 |