GradNorm (ResNetv2-101)

Reported on 3 benchmarks across 1 task · 1 paper · 3 SOTA

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

Computer Vision3 results

Out-of-Distribution DetectiononImageNet-1k vs Textures
FPR95· 2021-10-01
61.42
best: 79.8 (RP+GradNorm)
SOTA
On the Importance of Gradients for Detecting Distributional Shifts in the Wild arXiv:2110.00218
Out-of-Distribution DetectiononImageNet-1k vs Places
FPR95· 2021-10-01
60.86
best: 83.29 (RP+GradNorm)
SOTA
On the Importance of Gradients for Detecting Distributional Shifts in the Wild arXiv:2110.00218
Out-of-Distribution DetectiononImageNet-1k vs SUN
FPR95· 2021-10-01
46.48
best: 80.94 (DOE)
SOTA
On the Importance of Gradients for Detecting Distributional Shifts in the Wild arXiv:2110.00218