Image Classification on Tiny-Imagenet (Alpha-Weighted Accuracy)
30.85Accuracy (alpha=0.1)Ensemble
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| EnsembleCommunication rounds=12024.10 | 30.85 | 39.43 | 45.8 | |
| FuseFLK (number of splitted blocks)=2, Communication rounds=22024.10 | 29.28 | 31.11 | 34.34 | |
| FuseFLK (number of splitted blocks)=4, Communication rounds=42024.10 | 27.63 | 33.04 | 34.28 | |
| FuseFLK (number of splitted blocks)=8, Communication rounds=82024.10 | 25.21 | 32.59 | 33.82 | |
| DENSECommunication rounds=12024.10 | 22.44 | 28.14 | 32.34 | |
| Fed-ADICommunication rounds=12024.10 | 19.59 | 25.34 | 30.21 | |
| Fed-DAFLCommunication rounds=12024.10 | 18.38 | 22.18 | 28.22 | |
| FedDFCommunication rounds=12024.10 | 15.34 | 18.22 | 27.43 | |
| FedAvgCommunication rounds=12024.10 | 3.12 | 10.46 | 11.89 |