Transferability Estimation on NeuCRaB N/K = 10 benchmark leaderboard · SOTA2 Research

Transferability Estimation on NeuCRaB N/K = 10

0.276Estimation Error (100+ classes)LEEP

Evaluation Results

Method	Links
LEEP2022.03		0.276	0.079	0.049	0.134
LINEAR2022.03		0.325	0.089	0.109	0.174
H-score2022.03		0.461	0.318	0.158	0.313
LogME2022.03		0.488	0.138	0.073	0.233
PACTran-Gaussian2022.03		0.769	0.678	0.429	0.625
PACTran-Gaussian2022.03		0.778	0.609	0.534	0.641
N-LEEP2022.03		0.822	0.52	0.148	0.497
LINEAR-VALID2022.03		0.835	0.482	0.123	0.48
PACTran-Dirichlet2022.03		0.839	0.446	0.134	0.473
PACTran-Gamma2022.03		0.839	0.452	0.14	0.477