Closed Chokurei closed 8 years ago
I misunderstand what you mean just now, thank you so much MaxPooling2D((3, 3), strides=(2,2))
Hi, thank you for your question.
Indeed, there is overlapping in some MaxPooling layers. These aren't choices I have made myself, it is the setting of the original paper. There is a section in the paper about this overlapping choice (Section 3.4), maybe you will find more informations there : https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf
Thank you for the reply, I understand it now
You are welcome, I close the issue.
Thank you for your kind upload I tried your Alexnet, it seems nice
But how about the Overlapping pooling In your code, maxpooling stride is (3, 3), I think is (2, 3) or something Hope your reply
thank you a lot