Open Set Domain Generalization on Digits-DG MNIST-M (test)
74.5AccuracyMixStyle
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MixStyleBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 74.5 | 59.3 | 56.95 | |
| CIRLBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 73.78 | 59.88 | 58.25 | |
| MEDIC++Backbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 73.14 | 60.51 | 61.62 | |
| SWADBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 73.09 | 53.35 | 55.94 | |
| FishBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 71.81 | 47.31 | 56.05 | |
| MEDICBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 71.14 | 60.98 | 58.28 | |
| MLDGBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 71.11 | 46.84 | 55.17 | |
| OpenMaxBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 71.03 | 57.26 | — | |
| ERMBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 71.03 | 53.92 | 54.04 | |
| ARPLBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 69.78 | 58.08 | 54.21 |