Table 3 The detailed network parameters of GRU
From: Detection method of absence seizures based on Resnet and bidirectional GRU
layer | The name of network layer | Activation function | Output size | Convolution kernel size | Filters | parameters |
|---|---|---|---|---|---|---|
RNN_input | – | 64 × 16 × 164 | – | – | – | |
1 | Reshape | – | 64 × 1024 | – | – | – |
2 | Bidirectional | tanh | 64 × 128 | 0 | 128 | 886,272 |
3 | Bidirectional_1 | tanh | 64 × 128 | 0 | 128 | 18,144 |
4 | Time_distributed | – | 64 × 128 | – | 128 | 16,512 |
5 | Time_distributed_1 | – | 64 × 1 | – | 1 | 129 |
6 | Output | RELU | 64 × 1 | – | – | – |