Different output of the size of Layer
zenght opened this issue · comments
Haitao Zeng commented
Hello, I want to know that the feature map output of each layer of your Resnet is different from the default output of pytorch.
My implementation
('input:', (1, 3, 224, 224)) (1, 256, 56, 56) (1, 512, 28, 28) (1, 1024, 14, 14) (1, 2048, 7, 7) ('output:', (1, 1000))
Yours
input: (1, 3, 224, 224)
(1, 64, 57, 57)
(1, 256, 57, 57)
(1, 512, 29, 29)
(1, 1024, 15, 15)
(1, 2048, 8, 8)
output: (1, 1000)