Image Classification on CIFAR-100 (test) (Beta-Weighted Accuracy)
49.75Accuracy (beta=0.1)OnDev-LCT-8/1
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OnDev-LCT-8/1Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 49.75 | 59.71 | 62.42 | |
| OnDev-LCT-4/1Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 47.51 | 58.85 | 61.79 | |
| OnDev-LCT-2/1Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 44.83 | 56.74 | 59.53 | |
| OnDev-LCT-1/1Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 42.64 | 55.23 | 58.07 | |
| CCT-4/2Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 39.44 | 50.43 | 52.76 | |
| CCT-2/2Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 35.21 | 47.99 | 50.5 | |
| ResNet-44Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 32.85 | 43.6 | 44.81 | |
| ResNet-32Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 31.34 | 41.33 | 42.68 | |
| ResNet-20Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 27.89 | 37.08 | 37.53 | |
| MobileNetv2/0.5Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 23.4 | 33.28 | 32.58 | |
| MobileNetv2/0.2Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 21.52 | 30.57 | 31.67 | |
| ViT-Lite-2/8Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 17.99 | 28.2 | 32.32 | |
| ViT-Lite-1/8Federated Learning algorithm=FedAvg, Number of clients=10, Evaluation protocol=Best of 3 runs2024.01 | 17.85 | 27.6 | 31.37 |