neouyghur
commented
4 years ago This is what I got.
Layer (type) Output Shape Param #
Input (1, 3, 512, 512) 0
Conv2D-1 (1, 64, 256, 256) 3136
Conv2D-2 (1, 32, 256, 256) 2080
Conv2D-3 (1, 32, 256, 256) 9248
Conv2D-4 (1, 32, 256, 256) 9248
LC-5 (1, 32, 256, 256) 0
MaxPool2D-6 (1, 64, 128, 128) 0
Conv2D-7 (1, 64, 128, 128) 36928
Conv2D-8 (1, 64, 128, 128) 36928
Residual-9 (1, 64, 128, 128) 0
Conv2D-10 (1, 64, 128, 128) 36928
Conv2D-11 (1, 64, 128, 128) 36928
Residual-12 (1, 64, 128, 128) 0
Conv2D-13 (1, 32, 128, 128) 2080
Conv2D-14 (1, 32, 128, 128) 9248
Conv2D-15 (1, 32, 128, 128) 9248
LC-16 (1, 32, 128, 128) 0
Conv2D-17 (1, 128, 64, 64) 73856
Conv2D-18 (1, 128, 64, 64) 147584
Conv2D-19 (1, 128, 64, 64) 8320
Residual-20 (1, 128, 64, 64) 0
Conv2D-21 (1, 128, 64, 64) 147584
Conv2D-22 (1, 128, 64, 64) 147584
Residual-23 (1, 128, 64, 64) 0
Conv2D-24 (1, 64, 64, 64) 8256
Conv2D-25 (1, 64, 64, 64) 36928
Conv2D-26 (1, 64, 64, 64) 36928
LC-27 (1, 64, 64, 64) 0
Conv2D-28 (1, 256, 32, 32) 295168
Conv2D-29 (1, 256, 32, 32) 590080
Conv2D-30 (1, 256, 32, 32) 33024
Residual-31 (1, 256, 32, 32) 0
Conv2D-32 (1, 256, 32, 32) 590080
Conv2D-33 (1, 256, 32, 32) 590080
Residual-34 (1, 256, 32, 32) 0
Conv2D-35 (1, 128, 32, 32) 32896
Conv2D-36 (1, 128, 32, 32) 147584
Conv2D-37 (1, 128, 32, 32) 147584
LC-38 (1, 128, 32, 32) 0
Conv2D-39 (1, 512, 16, 16) 1180160
Conv2D-40 (1, 512, 16, 16) 2359808
Conv2D-41 (1, 512, 16, 16) 131584
Residual-42 (1, 512, 16, 16) 0
Conv2D-43 (1, 512, 16, 16) 2359808
Conv2D-44 (1, 512, 16, 16) 2359808
Residual-45 (1, 512, 16, 16) 0
Conv2D-46 (1, 2, 16, 16) 1026
Conv2D-47 (1, 256, 32, 32) 33024
Conv2DTranspose-48 (1, 256, 32, 32) 8448
ELU-49 (1, 256, 32, 32) 0
ELU-50 (1, 256, 32, 32) 0
Conv2D-51 (1, 128, 64, 64) 8320
Conv2DTranspose-52 (1, 128, 64, 64) 524416
ELU-53 (1, 128, 64, 64) 0
ELU-54 (1, 128, 64, 64) 0
Conv2D-55 (1, 64, 128, 128) 2112
Conv2DTranspose-56 (1, 64, 128, 128) 131136
ELU-57 (1, 64, 128, 128) 0
ELU-58 (1, 64, 128, 128) 0
Conv2D-59 (1, 3, 128, 128) 192
Conv2D-60 (1, 64, 256, 256) 2112
Conv2DTranspose-61 (1, 64, 256, 256) 65600
ELU-62 (1, 64, 256, 256) 0
ELU-63 (1, 64, 256, 256) 0
Conv2D-64 (1, 3, 256, 256) 192
Conv2DTranspose-65 (1, 3, 512, 512) 3075
ELU-66 (1, 3, 512, 512) 0
STE-67 (1, 256, 32, 32), (1, 128, 64, 64), (1, 3, 128, 128), (1, 3, 256, 256), (1, 3, 512, 512) 0
Parameters in forward computation graph, duplicate included
Total params: 12396357
Trainable params: 12396357
Non-trainable params: 0
Shared params in forward computation graph: 0
Unique parameters in model: 12396357
Hi, I would like to compare my network parameters with yours? Could you report the total number of your parameters? Thanks.