Skip to main content
. Author manuscript; available in PMC: 2020 Aug 12.
Published in final edited form as: IEEE Access. 2020;8:106296–106309. doi: 10.1109/access.2020.2999546

TABLE 4.

Comparison of the architecture of scaling networks. The layers considered, the width of the each layer and the total number of parameters is shown.

CRNN
architecture		 No. of parameters
3M 9M 15M 27M
Layer Type Layer Width
2D convolution 64 64 128 256
2D convolution 64 64 128 256
stacked GRU 70 100 150 150
Fully connected 256 512 512 1024