The recommended solution would be to use device=cuda instead of device=gpu to use the gpuarray backend. Also, you might want to use Theano 1.0 since a number of fixes have been made in the GPU backend.
If that isn’t possible, upgrade the nvidia driver. > Le 8 nov. 2018 à 03:19, limoran...@gmail.com a écrit : > > Hello, everyone, here is my problem. > > My environment is > Titan V; nvidia driver 390.77 > Ubuntu16.06; Cuda 8.0; cudnn 6.0.21; python2.7 anaconda > (I have also tried on ubuntu14.04 cuda7.5 cudnn5.1. Same error as below.) > And I used conda to install theano as: $conda install theano==0.9.0 > I configured ~/.theanorc as follows: > <theanorc.png> <about:invalid#zClosurez><Excption.png> > > > Thanks a lot for any replication!!! > > -- > > --- > 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 theano-users+unsubscr...@googlegroups.com > <mailto:theano-users+unsubscr...@googlegroups.com>. > For more options, visit https://groups.google.com/d/optout > <https://groups.google.com/d/optout>. > <theanorc.png><Excption.png> -- --- 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 theano-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.