Image Classification on DTD Pathological Non-IID (test)
87.67AccuracyFedOTP
Evaluation results
| Method | Links | |
|---|---|---|
| FedOTPBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 87.67 | |
| CoOpBackbone=ResNet-50, Federated Learning Setting=Local Training, Number of Clients=102024.02 | 82.69 | |
| pFedPromptBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 77.14 | |
| PromptFL+FTBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 53.74 | |
| PromptFL+FedProxBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 50.89 | |
| PromptFLBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 50.46 | |
| PromptFL+FedPerBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 50.23 | |
| PromptFL+FedAMPBackbone=ResNet-50, Federated Learning Setting=Pathological Non-IID, Number of Clients=102024.02 | 47.16 | |
| Zero-Shot CLIPBackbone=ResNet-50, Federated Learning Setting=Local Training (Zero-shot), Number of Clients=102024.02 | 40.21 |