Image Classification on CIFAR-10 (i.i.d.)
0.96SSIMFedSGD
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FedSGDBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1]2025.12 | 0.96 | — | 471.96 | — | |
| QSGDBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], Quantization=8-bit2025.12 | 0.93 | 12.97 | 117.98 | — | |
| Top-kBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], k=6002025.12 | 0.91 | 38.11 | 2.3 | — | |
| FedSGD + LaplaceBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], var=0.12025.12 | 0.84 | — | 471.96 | — | |
| SA-FedLoraBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1]2025.12 | 0.83 | 35.84 | 2.1 | — | |
| lp-projBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1]2025.12 | 0.74 | 34.72 | 1.84 | — | |
| FedSGD + LaplaceBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], var=102025.12 | 0.23 | — | — | — | |
| FedMPDDBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], m=2000, Compression=0.6 % d2025.12 | 0.22 | 36.26 | 3.26 | — | |
| FedMPDDBytes Budget (GB)=0.90, Target Acc (%)=60, Model Architecture=CNN model from [1], m=600, Compression=0.2 % d2025.12 | 0.14 | 40.84 | 1.32 | — |