Skip to main content
. Author manuscript; available in PMC: 2021 Oct 20.
Published in final edited form as: J Neurosci Methods. 2021 Jul 6;361:109282. doi: 10.1016/j.jneumeth.2021.109282

Table 3.

Architecture and parameters of the CRNN model

Layer Number of filters Kernel size Stride Output shape Number of parameters Regularization

Input - - - (1000, 27) 0 -
1D convolution 32 3 1 (998, 32) 2624 -
Max-pooling 32 3 2 (499, 32) 0 Dropout (0.5)
1D convolution 64 3 1 (497, 64) 6208 -
Max-pooling 64 3 2 (248, 64) 0 Dropout (0.5)
GRU 35 - - (248, 35) 10500 -
Time distributed 35 - - (248, 35) 1260 -
Global average pooling 35 - - (1, 35) 0 -
Dense 35 - - (1, 2) 72 -
Total 20664