Advances, Systems and Applications
Table 1 CNN model parameters
From: A new dynamic security defense system based on TCP_REPAIR and deep learning
Layer | Operation | Input | Kernel | Stride | Pad | Activation Function | Pooling Type | Pooling Size/Stride/Padding | Output |
|---|---|---|---|---|---|---|---|---|---|
1 | CONV | 784*1 | 25*1 | 1 | 12 | ReLU | Max | 3*1/3/1 | 262*32 |
2 | CONV | 262*32 | 25*1 | 1 | 12 | ReLU | Max | 3*1/3/1 | 88*64 |
3 | Full Connect | 88*64 | – | – | – | – | – | – | 1024 |
4 | Full Connect | 1024 | – | – | – | – | – | – | 2 |
5 | Softmax | 2 | – | – | – | – | – | – | 2 |