-
Notifications
You must be signed in to change notification settings - Fork 0
MaxPooling2D
Yi Zhu edited this page May 14, 2021
·
2 revisions
MaxPooling2D
is a layer that executes 2-dimensional maxpooling on the input data.
MaxPooling2D(;input_shape::Tuple, pool_size::Tuple{Int64, Int64}, strides::Tuple{Int64,Int64}=(pool_size[1],pool_size[2]))
input_shape
: a tuple of input data shape, (s1, s2, channels)
, can be set automatically, unless it is the first layer in the network
pool_size
: the shape of kernel, a tuple of two integers
strides
: the moving step size in vertical and horizontal direction, default (pool_size[1],pool_size[2])
Star the repo if you like it! :-)