Ensemble-of-Leave-out-Classifiers
Ensemble-of-Leave-out-Classifiers copied to clipboard
Comparison to ODIN
Hi, thanks for this helpful code. I have a question about the paper baseline. In ODIN implementation https://github.com/facebookresearch/odin there are results that are different from the one that are presented in the paper.
For instance: DenseNet CIFAR100 and TINc as OOD: AUROC of 97.1 in the paper vs ODIN's github of 94.5. Have you succeed in replicating one of the results?
Thanks a lot, Koby