That is the problem. You use the old interface. Use the new one and filter_dilation will work. We should not change the interface again, so this problem shouldn't happen again. The old interface was specific to one implementation, while the new one isn't.
Fred On Wed, Jan 25, 2017 at 9:21 AM, Jose Dolz <[email protected]> wrote: > Hi Fred, > > thank you for your answer. > > So far I use this conv: > > theano.tensor.nnet.conv3d2d.conv3d > > Jose > > On Tuesday, 24 January 2017 16:53:47 UTC-5, nouiz wrote: >> >> theano.tensor.nnet.abstract_conv.conv3d() support the dilation >> parameter. which conv3d() do you talk about? >> >> I suggest you to use the path I gave. This is the recommanded way. >> >> Fred >> >> On Tue, Jan 24, 2017 at 4:23 PM, Jose Dolz <[email protected]> wrote: >> >>> Hi, >>> >>> I have seen that since some months ago dilation convolutions are >>> possible for 2D kernels, by including one of the parameters of conv2d ( >>> >>> *filter_dilation)*I was wondering whether someone knows if there are or >>> will be some implementations of these dilation convolutions in 3D, because >>> the 3D version of conv2d does not support these type of convolutions. >>> >>> Thanks in advance!! >>> >>> -- >>> >>> --- >>> 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. > -- --- 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.
