Machine Translation on WMT EN-ET (test)
12.6BLEUMix-MoE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Mix-MoEBase Model=Llama3.2-1B, Training Strategy=Mix-MoE2026.05 | 12.6 | 25.3 | 63.8 | |
| MOE-LPRBase Model=Llama3.2-1B, Training Strategy=MOE-LPR2026.05 | 10.2 | 20.3 | 61.4 | |
| MoEBase Model=Llama3.2-1B, Training Strategy=MoE2026.05 | 8.6 | 16.9 | 57.2 | |
| Dense PT+SFTBase Model=Llama3.2-1B, Training Strategy=Dense PT+SFT2026.05 | 7 | 14.5 | 51 | |
| Full FinetuneBase Model=Llama3.2-1B, Training Strategy=Full Finetune2026.05 | 6 | 11.9 | 47 | |
| Qwen2.5-1.5BBase Model=Qwen2.5-1.5B2026.05 | 5.1 | 8.2 | 40.8 | |
| Qwen2.5-0.5BBase Model=Qwen2.5-0.5B2026.05 | 4 | 5.1 | 36.7 | |
| LLAMA-ProBase Model=Llama3.2-1B, Training Strategy=LLAMA-Pro2026.05 | 4 | 7.4 | 42.7 | |
| LoRABase Model=Llama3.2-1B, Training Strategy=LoRA2026.05 | 3.1 | 5.2 | 39.5 | |
| LLAMA-MOE v2Base Model=Llama3.2-1B, Training Strategy=LLAMA-MOE v22026.05 | 2.9 | 4.8 | 44.3 | |
| Llama3.2-1BBase Model=Llama3.2-1B2026.05 | 2.5 | 3.3 | 38.2 |