search

AITTSMD / MTCNN-Tensorflow

Reproduce MTCNN using Tensorflow
1.51k stars 713 forks source link

R-Net MaxPooling Padding method #181

Open waquey opened 6 years ago

waquey commented 6 years ago

Hello, I've found there are 2 maxPooling layers in R_Net.

net = slim.max_pool2d(net, kernel_size=[3, 3], stride=2, scope="pool1", padding='SAME')

net = slim.conv2d(net,num_outputs=48,kernel_size=[3,3],stride=1,scope="conv2")

net = slim.max_pool2d(net,kernel_size=[3,3],stride=2,scope="pool2")

However, the first one used padding='SAME' while the second one applied the defulat 'Valid' Is there any reason?

Will it work if we set these 2 pooling layers as padding='Valid' ?

Thanks

AITTSMD commented 6 years ago

@waquey Nothing special，thank u~