Image Classification on FEMNIST 10 new writers (test)
88.18Overall AccuracySGD + Lrank
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SGD + LrankBackbone=LeNet2023.11 | 88.18 | 93.29 | 93.51 | 64 | |
| SGDBackbone=LeNet2023.11 | 87.85 | 92.87 | 93.09 | 64 | |
| LOLSGDBackbone=LeNet2023.11 | 87.47 | 91.87 | 92.23 | 66.27 | |
| SGD + LdistillBackbone=LeNet2023.11 | 87.44 | 92.46 | 92.81 | 61.91 | |
| LOLSGD + LrankBackbone=LeNet2023.11 | 87.16 | 91.61 | 91.97 | 66.27 | |
| LOLSGD + LdistillBackbone=LeNet2023.11 | 85.76 | 90.34 | 91.25 | 63.56 | |
| LOLSGD + Ldistill + SEBackbone=LeNet2023.11 | 85.1 | 89.52 | 90.58 | 63.56 | |
| SourceBackbone=LeNet2023.11 | 84.67 | 88.67 | 89.6 | 64.99 | |
| Naive TargetBackbone=LeNet2023.11 | 82.67 | 92.07 | 92.07 | 35.6 |