Image Classification on Dogs & Cats (test)
80.98Accuracy (BC)SelecMix
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SelecMixCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 80.98 | 29.03 | 55 | |
| LfF+BECheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 74.91 | 39 | 56.96 | |
| BiasEditCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 69.04 | 94.37 | 81.71 | |
| BiasEditCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 59.04 | 96.91 | 77.98 | |
| LfFCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 54.43 | 59.75 | 57.09 | |
| DFACheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 45.95 | 73.36 | 59.66 | |
| LfF+BECheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 40.12 | 71.62 | 55.87 | |
| SoftConCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 33.06 | 85.68 | 59.37 | |
| SelecMixCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 32.38 | 72.3 | 52.34 | |
| LfFCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 30.35 | 83.45 | 56.9 | |
| BiasAdvCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 25.91 | 88.33 | 57.12 | |
| VanillaCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 25.83 | 93.51 | 59.67 | |
| DFACheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 23.81 | 91.2 | 57.29 | |
| SoftConCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 23.63 | 97.81 | 60.72 | |
| DeNetDMCheckpoint selection=best bias-conflict test-set performance, Training bias-conflict ratio=0%, Number of runs=32026.05 | 23 | 93.91 | 58.46 | |
| VanillaCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 20.03 | 97.32 | 58.67 | |
| DeNetDMCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 17.04 | 95.37 | 56.21 | |
| BiasAdvCheckpoint selection strategy=best in-distribution validation performance, Training bias-conflict ratio=0%2026.05 | 15.41 | 95.12 | 55.27 |