4d tensor. I would suggest that you don't use manually dnn_conv. Let Theano optimizer handle that. use theano.tensor.nnet.conv2d()
Fred On Tue, Apr 4, 2017 at 1:50 AM Varglur <[email protected]> wrote: > hey, > > for the dnn_conv > <http://deeplearning.net/software/theano/library/sandbox/cuda/dnn.html#theano.sandbox.cuda.dnn.dnn_conv> > in theano , > about the parameter of *kerns*, how to define it ? 4D tensors ? or 3D > tensors ? > > > > - *img* – Images to do the convolution over. > - *kerns* – Convolution filters. > > -- > > --- > 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.
