Movement Regression on FALCON-M1
81.95R2 ScoreTEACHER
Evaluation results
| Method | Links | |
|---|---|---|
| TEACHERBackbone=NDT22026.01 | 81.95 | |
| IND +TSKDDistillation Strategy=Task-Specific Knowledge Distillation2026.01 | 43.52 | |
| IND +RDIMKDDistillation Strategy=Relationship Distillation2026.01 | 41.57 | |
| IND +KDDistillation Strategy=Knowledge Distillation2026.01 | 41.46 | |
| IND +VKDDistillation Strategy=Variable KD2026.01 | 40.73 | |
| IND +TOFDDistillation Strategy=Task-Oriented Feature Distillation2026.01 | 39.21 | |
| INDDistillation=None2026.01 | 37.59 | |
| IND +SIMKDDistillation Strategy=Similarity KD2026.01 | 37.3 | |
| IND +TEDDistillation Strategy=Task-aware Embedding Distillation2026.01 | 34.83 |