Table 1 The CNN layers and parameters.
Layer | Type | Output shape | Kernel | Stride | Param # |
|---|---|---|---|---|---|
1 | Input | (None, 726, 19) | – | – | 0 |
2 | Convolution 1D | (None, 695, 1) | 32 | 1 | 609 |
3 | Real-valued fast Fourier transform | (None, 348, 1) | – | – | 0 |
4 | Global maximum pooling | (None, 1) | – | – | 0 |
5 | Batch normalization | – | – | 2 | |
6 | Dense | (None, 2) | – | – | 4 |
Total (non-trainable) | 615 (2) |
