Machine Translation on WMT FI-EN (test)
15BLEUMix-MoE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Mix-MoEBase Model=Llama3.2-1B, Training Strategy=Mix-MoE2026.05 | 15 | 34.8 | 75.7 | |
| MOE-LPRBase Model=Llama3.2-1B, Training Strategy=MOE-LPR2026.05 | 14.2 | 33.8 | 74.9 | |
| MoEBase Model=Llama3.2-1B, Training Strategy=MoE2026.05 | 12.8 | 31.2 | 73.9 | |
| Full FinetuneBase Model=Llama3.2-1B, Training Strategy=Full Finetune2026.05 | 11.8 | 30.1 | 67.4 | |
| Qwen2.5-1.5BBase Model=Qwen2.5-1.5B2026.05 | 11.2 | 23.6 | 65.5 | |
| Dense PT+SFTBase Model=Llama3.2-1B, Training Strategy=Dense PT+SFT2026.05 | 10.5 | 31.5 | 69.5 | |
| LLAMA-ProBase Model=Llama3.2-1B, Training Strategy=LLAMA-Pro2026.05 | 9.4 | 26.6 | 62.7 | |
| LoRABase Model=Llama3.2-1B, Training Strategy=LoRA2026.05 | 9.3 | 26.7 | 65.6 | |
| LLAMA-MOE v2Base Model=Llama3.2-1B, Training Strategy=LLAMA-MOE v22026.05 | 9.2 | 25 | 73.9 | |
| Qwen2.5-0.5BBase Model=Qwen2.5-0.5B2026.05 | 4.3 | 13.5 | 55.1 | |
| Llama3.2-1BBase Model=Llama3.2-1B2026.05 | 3.8 | 7.5 | 52.1 |