Knowledge Distillation on MNIST + FASHION-MNIST
98.48AccuracyMDA
Evaluation results
| Method | Links | |
|---|---|---|
| MDABackbone=VGG-16, Category=0-2 + 0-22022.03 | 98.48 | |
| MDABackbone=VGG-16, Category=0-9 + 0-22022.03 | 98.29 | |
| Base (Source models average)Backbone=VGG-16, Category=0-9 + 0-22022.03 | 98.24 | |
| MDABackbone=VGG-16, Category=0+02022.03 | 97.35 | |
| Base (Source models average)Backbone=VGG-16, Category=0-2 + 0-22022.03 | 96.69 | |
| Base (Source models average)Backbone=VGG-16, Category=0-2 + 3-92022.03 | 96.12 | |
| MDABackbone=VGG-16, Category=0-2 + 3-92022.03 | 95.62 | |
| Base (Source models average)Backbone=VGG-16, Category=0+02022.03 | 94.6 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0-2 + 0-22022.03 | 93.81 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0-9 + 0-22022.03 | 93.57 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0-2 + 3-92022.03 | 92.76 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0+02022.03 | 92.24 |