Image Classification on NICO++ Feature non-IID
92.91Top-1 AccuracyHyperLoRA
Evaluation results
| Method | Links | |
|---|---|---|
| HyperLoRABackbone=ViT2026.06 | 92.91 | |
| CentralizedBackbone=ViT2026.06 | 91.51 | |
| LoRA-FAIRBackbone=ViT2026.06 | 91.24 | |
| FLoRABackbone=ViT2026.06 | 90.93 | |
| FlexLoRABackbone=ViT2026.06 | 90.6 | |
| FedITBackbone=ViT2026.06 | 90.58 | |
| FFA-LoRABackbone=ViT2026.06 | 90.28 | |
| HyperLoRABackbone=MLP-Mixer2026.06 | 85.86 | |
| CentralizedBackbone=MLP-Mixer2026.06 | 84.5 | |
| LoRA-FAIRBackbone=MLP-Mixer2026.06 | 83.56 | |
| FlexLoRABackbone=MLP-Mixer2026.06 | 83.08 | |
| FedITBackbone=MLP-Mixer2026.06 | 82.51 | |
| FLoRABackbone=MLP-Mixer2026.06 | 82.29 | |
| FFA-LoRABackbone=MLP-Mixer2026.06 | 80.05 |