Image Classification on Office-Home (Art Source Transfers)
61.22Accuracy (A->C)FedSCAl
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FedSCAlInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 61.22 | 87.16 | 88.76 | 79.05 | |
| FedProxInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 57.95 | 83.48 | 86.26 | 75.9 | |
| FedLoAInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 56.69 | 83.56 | 86.38 | 75.55 | |
| FedMOONInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 56.23 | 85.08 | 87.13 | 76.15 | |
| FedWCAInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 53 | 84.02 | 86.72 | 74.58 | |
| Dual AdaptInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 45.85 | 76.56 | 84.11 | 68.84 | |
| LoAInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 44.95 | 79.75 | 84.86 | 69.85 | |
| LADDInitial Server Model Pre-training Domain=Art, Backbone=ViT small, Pre-training Dataset=ImageNet, Participation Rate=0.3, Local Training Epochs=52025.12 | 37.5 | 65.21 | 71.23 | 57.98 |