To use cudnn 5 and 5.1, you must use Theano dev version. Theano 0.8.2 don't support it. This is probably your problem.
If not, then give us the full error message. Frédéric On Thu, Feb 2, 2017 at 4:14 PM Feras Almasri <[email protected]> wrote: > After I changed the pooling to sum I run into this problem > > poolsize = (image_shape[0], image_shape[1]) > pooled_out = pool.pool_2d( > input= self.input, > ds=poolsize, > ignore_border=True, > mode= 'sum' > ) > self.output = pooled_out > > ERROR (theano.gof.opt): Optimization failure due to: > local_pool_dnn_alternative > ERROR (theano.gof.opt): node: Pool{ds=(20, 20), ignore_border=True, > st=(20, 20), padding=(0, 0), mode='sum'}(HostFromGpu.0) > > Knowing this: > Using gpu device 0: TITAN X (Pascal) (CNMeM is disabled, cuDNN 5105) > UserWarning: Your cuDNN version is more recent than the one Theano > officially supports. If you see any problems, try updating Theano or > downgrading cuDNN to version 5. > warnings.warn(warn) > > > -- > > --- > 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.
