Vision-Language Model Selection on 21 Datasets (Few-Shot)
0.707NDCG@5DCD
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DCDTarget images (Ntgt)=1, Source images per task (Nsrc)=25, Number of random seeds=102026.02 | 0.707 | 0.365 | 1.072 | 0.402 | 0.752 | |
| SWABNumber of random seeds=102026.02 | 0.616 | 0.318 | 0.934 | 0.393 | 0.691 | |
| ModelGPTNumber of random seeds=102026.02 | 0.544 | 0.267 | 0.811 | 0.386 | 0.702 | |
| INB2026.02 | 0.539 | 0.268 | 0.807 | 0.384 | 0.651 | |
| AvgRank2026.02 | 0.524 | 0.254 | 0.778 | 0.386 | 0.699 |