Table 1.
3Network parameters | LeNet-300-100 | LeNet-5 |
---|---|---|
Total parameters including weights and bias | 266610 | 44426 |
Learning rate | 0.001 | 0.001 |
Weight decay | 0.0001 | 0.0001 |
Sensitivity value (ρ) | 2.3 | 2.4 |
Training dataset size (MNIST) | 50000 | 50000 |
Test dataset size (MNIST) | 10000 | 10000 |
Training dataset size (CIFAR-10) | 50000 | 50000 |
Test dataset size (CIFAR-10) | 10000 | 10000 |