Implementation of the paper Distilling the Knowledge in a Neural Network
Results on MNIST
Test set size = 10000
Model | Accuracy |
---|---|
Teacher model | 0.9847 |
Distilled student model | 0.9810 |
Compression ratio = 2
Implementation of the paper Distilling the Knowledge in a Neural Network
Results on MNIST
Test set size = 10000
Model | Accuracy |
---|---|
Teacher model | 0.9847 |
Distilled student model | 0.9810 |
Compression ratio = 2