Transferability Estimation on CIFAR-10
0.949Weighted Kendall's TauSFDA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SFDAWall-Clock Time (s)=576.6, Backbone=11 Supervised CNN models, Pre-trained=ImageNet2022.07 | 0.949 | — | |
| LogMEWall-Clock Time (s)=23.5, Backbone=11 Supervised CNN models, Pre-trained=ImageNet2022.07 | 0.852 | — | |
| LEEPWall-Clock Time (s)=22.3, Backbone=11 Supervised CNN models, Pre-trained=ImageNet2022.07 | 0.824 | — | |
| NLEEPWall-Clock Time (s)=11000, Backbone=11 Supervised CNN models, Pre-trained=ImageNet2022.07 | 0.806 | — | |
| SFDA2022.07 | 0.748 | — | |
| SFDAModel type=ViT models2022.07 | 0.743 | 121.7 | |
| LogMEModel type=ViT models2022.07 | 0.741 | 13.1 | |
| LogME2022.07 | 0.687 | — | |
| NLEEPModel type=ViT models2022.07 | 0.674 | 8,000 | |
| SFDAWall-Clock Time (s)=504.12022.07 | 0.619 | — | |
| NLEEP2022.07 | 0.491 | — | |
| LogMEWall-Clock Time (s)=89.32022.07 | 0.295 | — | |
| NLEEPWall-Clock Time (s)=80002022.07 | -0.044 | — |