Open Set Domain Generalization on Digits-DG MNIST (test)
98.44AccuracyMEDIC++
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MEDIC++Backbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 98.44 | 80.58 | 96.88 | |
| CIRLBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.92 | 81.14 | 93.5 | |
| MEDICBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.89 | 83.2 | 95.81 | |
| MixStyleBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.86 | 73.25 | 89.36 | |
| MLDGBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.83 | 80.36 | 94.28 | |
| FishBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.83 | 74.69 | 95.61 | |
| ARPLBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.75 | 85.74 | 91.86 | |
| SWADBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.71 | 84.44 | 92.65 | |
| ERMBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.47 | 80.9 | 92.6 | |
| OpenMaxBackbone=ConvNet, Known:Unknown Ratio=6:42026.06 | 97.33 | 52.03 | — |