Image Classification on Fashion-MNIST Non-IID Dir(0.3) 2017 (test)
89.31Top-1 AccuracyFedNova
Evaluation results
| Method | Links | |
|---|---|---|
| FedNovaScale=0.75×2026.03 | 89.31 | |
| SFedHIFIScale=1.0×2026.03 | 89.18 | |
| SFedHIFIScale=0.75×2026.03 | 88.99 | |
| FedProxScale=0.75×2026.03 | 88.91 | |
| FedProxScale=0.5×2026.03 | 88.7 | |
| FedNovaScale=1.0×2026.03 | 88.63 | |
| SFedHIFIScale=Avg.2026.03 | 88.43 | |
| FedNovaScale=Avg.2026.03 | 88.41 | |
| SFedHIFIScale=0.5×2026.03 | 88.31 | |
| FedProxScale=1.0×2026.03 | 88.29 | |
| FedProxScale=Avg.2026.03 | 88.04 | |
| FedAVGScale=1.0×2026.03 | 87.98 | |
| FedNovaScale=0.5×2026.03 | 87.97 | |
| FedNovaScale=0.25×2026.03 | 87.74 | |
| SFedHIFIScale=0.25×2026.03 | 87.27 | |
| FedProxScale=0.25×2026.03 | 86.24 | |
| FedAVGScale=0.75×2026.03 | 86.14 | |
| FedAVGScale=Avg.2026.03 | 85.33 | |
| FedAVGScale=0.5×2026.03 | 83.86 | |
| FedAVGScale=0.25×2026.03 | 83.35 |