if i give something like this to maxpool2d
F.max_pool2d(self.conv1(x), 2))
the 2 here defines the stride, just to clarify
what would the kernel size be?
if i give something like this to maxpool2d
F.max_pool2d(self.conv1(x), 2))
the 2 here defines the stride, just to clarify
what would the kernel size be?
You define the kernel_size
in the call. If no stride
was set, it will be set to the kernel_size
as well.
but in this call what would the kernel size and stride be, both 2?
Yes, in your example code both values will be set to 2
, such that the spatial dimensions will be halved.