Table 3 Architecture of the GRU layers.
From: Iterative multiblock framework for high frequency EEG based neurological disorder detection
Layers | Kernel size | Filters | Trainable parameters |
|---|---|---|---|
Conv_1 | 3 | 64 | 640 |
batch_normalization | ----- | ------- | 256 |
Conv_1 | 5 | 64 | 1664 |
batch_normalization | ----- | ------- | 256 |
Conv_1 | 7 | 64 | 3200 |
batch_normalization | ----- | ------- | 256 |
concatenate | |||
max_pooling2d | 2 | --- | --- |
flatten | --- | ---- | --- |
GRU (bidirectional) | --- | 128 | 99,840 |
Attention | --- | ----- | |
dense | --- | ----- | 16,448 |
dropout | --- | ----- | |
dense_1 | --- | ----- | 130 |
