Image Classification on NICO++ Feature+label non-IID (average across domains)
91.51Top-1 AccuracyCentralized
Evaluation results
| Method | Links | |
|---|---|---|
| CentralizedBackbone=ViT2026.06 | 91.51 | |
| HyperLoRABackbone=ViT2026.06 | 91.19 | |
| LoRA-FAIRBackbone=ViT2026.06 | 90.04 | |
| FlexLoRABackbone=ViT2026.06 | 89.65 | |
| FLoRABackbone=ViT2026.06 | 89.6 | |
| FedITBackbone=ViT2026.06 | 89.48 | |
| FFA-LoRABackbone=ViT2026.06 | 89.45 | |
| CentralizedBackbone=MLP-Mixer2026.06 | 84.5 | |
| HyperLoRABackbone=MLP-Mixer2026.06 | 81.15 | |
| LoRA-FAIRBackbone=MLP-Mixer2026.06 | 79.53 | |
| FlexLoRABackbone=MLP-Mixer2026.06 | 78.73 | |
| FedITBackbone=MLP-Mixer2026.06 | 78.53 | |
| FLoRABackbone=MLP-Mixer2026.06 | 78.41 | |
| FFA-LoRABackbone=MLP-Mixer2026.06 | 76.78 |