List your company

Out-of-Distribution Detection on SUN (test)

40.63FPR@95MOS
38.320453.910269.585.0898May 5, 2021Jul 27, 2021Oct 19, 2021Jan 10, 2022Apr 4, 2022Jun 26, 2022Sep 18, 2022

Evaluation results

MethodLinks
MOSBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=3.2, Fine-tuned=true2021.0540.6392.01
RankFeatBackbone=SqueezeNet, Block=Block 3 + 42022.0946.6490.17
RankFeatBackbone=SqueezeNet, Block=Block 42022.0946.7288.31
KNN (non-parametric)Backbone=ViT-B/16, Fine-tuned on=ImageNet-1k2022.0448.4—
RankFeatBackbone=SqueezeNet, Block=Block 32022.0949.1890.18
GradNormBackbone=SqueezeNet2022.0953.6387.55
EnergyBackbone=SqueezeNet2022.0956.4187.88
RankFeatBackbone=T2T-ViT-242022.0957.1884.33
MSPBackbone=T2T-ViT-242022.0961.7781.37
ODINBackbone=T2T-ViT-242022.0963.8378.46
ReActBackbone=T2T-ViT-242022.0965.2381.03
EnergyBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=3.1, Fine-tuned=true2021.0565.3385.32
KL MatchingBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=20.6, Fine-tuned=true2021.0567.5278.72
EnergyBackbone=T2T-ViT-242022.0968.5573.06
ODINBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=23.6, Fine-tuned=true2021.0571.6783.92
ODINBackbone=SqueezeNet2022.0978.3278.37
MSPBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=3.1, Fine-tuned=true2021.0579.9878.34
Mahalanobis (parametric)Backbone=ViT-B/16, Fine-tuned on=ImageNet-1k2022.0480.51—
MSPBackbone=SqueezeNet2022.0983.0372.25
ReActBackbone=SqueezeNet2022.0987.5766.37
MahalanobisBackbone=BiT-S-R101x1, In-distribution dataset=ImageNet-1k, Test Time (min)=145.4, Fine-tuned=true2021.0588.4365.2
MahalanobisBackbone=SqueezeNet2022.0990.3362.18
MahalanobisBackbone=T2T-ViT-242022.0991.7150.52
GradNormBackbone=T2T-ViT-242022.0998.3728.06