For uspample = 4, something like this:
output = T.zeros(out_size, dtype=theano.config.floatX)
for i in range(upsample):
for j in range(upsample):
output = T.set_subtensor(output[:, :,
i::upsample,j::upsample], input)
will give you
array([[[[ 1., 1., 1., 1., 2., 2., 2., 2.],
[ 1., 1., 1., 1., 2., 2., 2., 2.],
[ 1., 1., 1., 1., 2., 2., 2., 2.],
[ 1., 1., 1., 1., 2., 2., 2., 2.],
[ 3., 3., 3., 3., 4., 4., 4., 4.],
[ 3., 3., 3., 3., 4., 4., 4., 4.],
[ 3., 3., 3., 3., 4., 4., 4., 4.],
[ 3., 3., 3., 3., 4., 4., 4., 4.]]]])
for input array([[[[ 1., 2.],
[ 3., 4.]]]])
On Mon, Aug 29, 2016 at 7:29 AM, <[email protected]> wrote:
> how did you solve this problem,I meet this problem ,too
>
> 在 2014年12月14日星期日 UTC+8上午3:28:38，Mohammad Havaei写道：
>
>> Does there exist in Theano an operation to resize a matrix by
>> interpolation or downsample it?
>>
> --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "theano-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>
--
---
You received this message because you are subscribed to the Google Groups
"theano-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
