An Empirical Investigation of Pre-trained Model Selection for Out-of-Distribution Generalization and Calibration

Open in new window