Federated Image Classification on MNIST (test)
16Communication RoundsGradMA-S
Evaluation results
| Method | Links | |
|---|---|---|
| GradMA-Somega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 16 | |
| FedMLBMomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 16 | |
| GradMAomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 18 | |
| FedAvgMomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 19 | |
| MOONMomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 19 | |
| Feddynomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 19 | |
| FedMLBomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 22 | |
| FedProxMomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 22 | |
| FedAvgomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 25 | |
| FedProxomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 25 | |
| MOONomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 25 | |
| Scaffoldomega=1.0, S=10, target accuracy=95%, backbone=NN, communication load=2x per round2023.02 | 26 | |
| GradMAomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 29 | |
| GradMA-Womega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 30 | |
| MimeLiteomega=1.0, S=10, target accuracy=95%, backbone=NN, communication load=2x per round2023.02 | 32 | |
| GradMAomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 37 | |
| GradMA-Somega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 38 | |
| GradMA-Somega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 38 | |
| MIFAMomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 39 | |
| MIFAomega=1.0, S=10, target accuracy=95%, backbone=NN2023.02 | 42 | |
| GradMA-Somega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 43 | |
| GradMAomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 49 | |
| Scaffoldomega=0.01, S=5, target accuracy=45%, backbone=NN, communication load=2x per round2023.02 | 60 | |
| MIFAMomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 60 | |
| MIFAomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 62 | |
| MIFAMomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 62 | |
| MIFAomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 67 | |
| Feddynomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 73 | |
| GradMA-Womega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 78 | |
| MIFAMomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 80 | |
| MIFAomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 85 | |
| FedMLBMomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 87 | |
| GradMA-Womega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 111 | |
| FedMLBMomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 113 | |
| FedAvgomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 115 | |
| FedMLBomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 115 | |
| FedProxMomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 115 | |
| MOONMomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 115 | |
| MimeLiteomega=0.1, S=10, target accuracy=95%, backbone=NN, communication load=2x per round2023.02 | 115 | |
| MOONomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 116 | |
| FedAvgMomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 116 | |
| FedProxomega=0.1, S=10, target accuracy=95%, backbone=NN2023.02 | 120 | |
| GradMA-Womega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 125 | |
| Feddynomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 128 | |
| Feddynomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 136 | |
| FedAvgMomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 138 | |
| FedProxMomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 138 | |
| MOONMomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 186 | |
| FedMLBMomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 191 | |
| FedMLBomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 203 | |
| MOONMomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 210 | |
| MimeLiteomega=0.01, S=5, target accuracy=45%, backbone=NN, communication load=2x per round2023.02 | 277 | |
| FedMLBomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 280 | |
| FedAvgMomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 280 | |
| FedProxMomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 280 | |
| FedProxomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 283 | |
| MOONomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 283 | |
| MimeLiteomega=0.01, S=50, target accuracy=40%, backbone=NN, communication load=2x per round2023.02 | 284 | |
| FedAvgomega=0.01, S=50, target accuracy=40%, backbone=NN2023.02 | 299 | |
| FedProxomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 478 | |
| FedAvgomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 493 | |
| MOONomega=0.01, S=5, target accuracy=45%, backbone=NN2023.02 | 493 |