Table 4.
layer | type | parameters |
---|---|---|
1 | ConvLSTM | k = (n × 2 × 3), f = 16, s = 1 |
2 | ConvLSTM | k = (16 × 1 × 3), f = 32, s = (1 × 2) |
3 | ConvLSTM | k = (32 × 1 × 3), f = 64, s = (1 × 2) |
4 | FC | m = 256, sigmoid |
5 | FC | m = 2, sigmoid |
k, kernel dimensions; n, input channel dimensions; f, number of filters; s, stride; FC, fully connected layer; m, number of neurons.