MaxPool Problem about pipecnn HOT 3 CLOSED

Hi @AndrewSivrit if you are using pool_stride=1, regardless what pool_size you are using, you will always get the same output size after pooling, it is the nature of the algorithm, not determined by the code.

So what kind of pooling operation you want to implement ?

from pipecnn.

Doonny are you sure ?
If I have feature map 13x13 and pool_size = 2, pool_stride = 1, I get result (13-1)x(13-1) feature map, because in the line (size = 13) filter of pooling has just 12 unique positions.

from pipecnn.

@AndrewSivrit I am also trying to use feature map 13x13 with pool_size = 2 and pool_stride = 1, getting an error incorrect setting of pooling input/output size for layer-6!!!

How did you fix this?

from pipecnn.