Knowledge Distillation on CIFAR-10 + Tiny-ImageNet
94.7AccuracyMDA
Evaluation results
| Method | Links | |
|---|---|---|
| MDABackbone=VGG-16, Category=0+02022.03 | 94.7 | |
| Base (Source models average)Backbone=VGG-16, Category=0+02022.03 | 88.2 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0+02022.03 | 85.62 | |
| Base (Source models average)Backbone=VGG-16, Category=3-9 + 0-692022.03 | 54.02 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=3-9 + 0-692022.03 | 53.26 | |
| MDABackbone=VGG-16, Category=0-2 + 0-692022.03 | 53.2 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0-2 + 0-692022.03 | 52.67 | |
| Base (Source models average)Backbone=VGG-16, Category=0-2 + 0-692022.03 | 51.97 | |
| Knowledge Amalgamating (KA)Backbone=VGG-16, Category=0-9 + 70-1792022.03 | 51.84 | |
| MDABackbone=VGG-16, Category=3-9 + 0-692022.03 | 50.2 | |
| Base (Source models average)Backbone=VGG-16, Category=0-9 + 70-1792022.03 | 49.33 | |
| MDABackbone=VGG-16, Category=0-9 + 70-1792022.03 | 42.3 |