ViT-L/32-384 with Entropy

Reported on 3 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology3 results

ClassificationonImageNet C-OOD (class-out-of-distribution)
Detection AUROC (severity 0)· 2023-02-23
0.9948
best: 0.9958 (ViT-L/32-384 with Max-logit)
A framework for benchmarking class-out-of-distribution detection and its application to ImageNet arXiv:2302.11893
ClassificationonImageNet C-OOD (class-out-of-distribution)
Detection AUROC (severity 10)· 2023-02-23
0.7332
best: 0.7748 (ViT-L/32-384 with Max-logit)
A framework for benchmarking class-out-of-distribution detection and its application to ImageNet arXiv:2302.11893
ClassificationonImageNet C-OOD (class-out-of-distribution)
Detection AUROC (severity 5)· 2023-02-23
0.9514
best: 0.9632 (ViT-L/32-384 with Max-logit)
A framework for benchmarking class-out-of-distribution detection and its application to ImageNet arXiv:2302.11893