Dense Model Loss Curve Prediction Cosine to WSD transfer
0.404MAPECMMT (λ = 0.99)
Evaluation Results
| Method | Links | |
|---|---|---|
| CMMT (λ = 0.99)Model Size=100M2025.12 | 0.404 | |
| ASMTModel Size=100M2025.12 | 0.41 | |
| CMMT (λ = 0.999)Model Size=100M2025.12 | 0.422 | |
| CMMT (λ = 0.99)Model Size=500M2025.12 | 0.657 | |
| ASMTModel Size=50M2025.12 | 0.72 | |
| CMMT (λ = 0.99)Model Size=50M2025.12 | 0.777 | |
| CMMT (λ = 0.999)Model Size=500M2025.12 | 0.784 | |
| ASMTModel Size=500M2025.12 | 0.798 | |
| CMMT (λ = 0.999)Model Size=50M2025.12 | 0.838 |