Image Classification on CIFAR-10 (10%-100 devices)
83.56Accuracy (Dirichlet α=0.5)FedVSSAM
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| FedVSSAMDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 83.56 | 111 | 79.57 | 296 | 83.57 | 107 | 81.11 | 196 | |
| FedGAMMADevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 82.36 | 201 | 77.02 | 208 | 82.74 | 179 | 79.49 | 393 | |
| MoFedSAMDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 81.91 | 199 | 78.94 | 385 | 82.94 | 178 | 81.15 | 291 | |
| FedSMOODevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 81.49 | 147 | 78.85 | 358 | 82.87 | 134 | 80.76 | 289 | |
| SCAFFOLDDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 80.14 | 170 | 75.81 | 635 | 81.35 | 136 | 78.72 | 326 | |
| FedACGDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 78.84 | 113 | 76.85 | 457 | 80.62 | 85 | 79.02 | 199 | |
| FedSAMDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 78.75 | 266 | 74.06 | — | 80.25 | 240 | 76.07 | 639 | |
| FedAdamDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 78.35 | 191 | 75.76 | 741 | 78.13 | 188 | 77.26 | 440 | |
| FedAvgDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 77.05 | 315 | 72.57 | — | 77.55 | 299 | 74.63 | — | |
| FedLESAMDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 76.88 | 306 | 72.31 | — | 77.64 | 240 | 75.49 | 736 | |
| FedProxDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 76.82 | 337 | 72.23 | — | 77.81 | 271 | 74.89 | — | |
| FedDynDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 76.8 | 356 | 74.6 | — | 77.19 | 255 | 76.78 | 446 | |
| FedCMDevice Sampling=10%-100 devices, Training Rounds=8002026.05 | 76.23 | 388 | 74.07 | — | 77.48 | 212 | 76.14 | 574 |