Image Classification on PathMNIST (test)
99.63Accuracy@1 (%)CE
Evaluation results
| Method | Links | |
|---|---|---|
| CESamples per Class=full dataset2026.06 | 99.63 | |
| Distributional LossSamples per Class=full dataset, Density Estimate=nearest-neighbor2026.06 | 99.61 | |
| PolyLossSamples per Class=full dataset2026.06 | 99.53 | |
| T-vMFSamples per Class=full dataset2026.06 | 99.43 | |
| Distributional LossSamples per Class=200, Density Estimate=nearest-neighbor2026.06 | 96.06 | |
| T-vMFSamples per Class=2002026.06 | 95.44 | |
| PolyLossSamples per Class=2002026.06 | 95.04 | |
| CESamples per Class=2002026.06 | 94.72 | |
| Distributional LossSamples per Class=20, Density Estimate=nearest-neighbor2026.06 | 90.04 | |
| T-vMFSamples per Class=202026.06 | 88.28 | |
| CESamples per Class=202026.06 | 83.5 | |
| FedSOLNon-IID Partition Strategy=LDA, alpha=0.12023.08 | 78.88 | |
| PolyLossSamples per Class=202026.06 | 75.35 | |
| FedSOLNon-IID Partition Strategy=Sharding, s=22023.08 | 73.85 | |
| FedAvgNon-IID Partition Strategy=LDA, alpha=0.12023.08 | 65.98 | |
| PolyLossSamples per Class=22026.06 | 61.19 | |
| Distributional LossSamples per Class=2, Density Estimate=nearest-neighbor2026.06 | 59.76 | |
| FedAvgNon-IID Partition Strategy=Sharding, s=22023.08 | 57.4 | |
| T-vMFSamples per Class=22026.06 | 56.94 | |
| CESamples per Class=22026.06 | 56.29 |