Machine Translation on UM-Law
58.82BLEUFinetune
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Finetune2022.11 | 58.82 | 0.6375 | |
| PLACPruning Ratio=30%2022.11 | 58.65 | 0.6056 | |
| Adaptive kNNPruning Ratio=0%2022.11 | 58.64 | 0.6017 | |
| KnownPruning Ratio=30%2022.11 | 56.92 | 0.5762 | |
| MergePruning Ratio=30%2022.11 | 56.51 | 0.5873 | |
| RandomPruning Ratio=30%2022.11 | 53.78 | 0.5661 | |
| ClusterPruning Ratio=30%2022.11 | 49.65 | 0.5274 | |
| All KnownPruning Ratio=63%2022.11 | 46.45 | 0.472 | |
| Base2022.11 | 30.36 | 0.3857 |