Image Classification on CIFAR100 Pathological Non-IID
68.28AccuracyOracle
Evaluation results
| Method | Links | |
|---|---|---|
| OracleBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 68.28 | |
| pFedLIABackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%, Warm-up Rounds=202025.05 | 68.13 | |
| FeSEMBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 51.2 | |
| CFLBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 48.47 | |
| FedAvgBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 48.2 | |
| FedFomoBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 43.77 | |
| FedProtoBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 36.68 | |
| Local-onlyBackbone=ResNet-18, Pre-trained=true, Data Heterogeneity Setting=Pathological Non-IID, Client Participation Rate=10%2025.05 | 34.39 |