I'm sorry I want to correct myself. use_cuda =None is not creating a new process on another GPU and is running as usual. Only using use_cuda = True in libgpuarray mode has this issue.
On Saturday, November 12, 2016 at 7:46:17 PM UTC-7, Ragav Venkatesan wrote: > > same result from what I can see. But I was under the assumption that > use_cuda = None is going make the thing run on CPU no matter what. > > On Saturday, November 12, 2016 at 9:31:12 AM UTC-7, Pascal Lamblin wrote: >> >> What about the default of "use_cuda=None"? >> >> On Fri, Nov 11, 2016, Ragav Venkatesan wrote: >> > I use theano.sandbox.rng_mrg.MRG_RandomStreams to generate a 2D (and 1D >> ) >> > dropout masks for applying dropouts. I create it so, >> > >> > srng = MRG_RandomStreams(rng.randint(1,89324723894), use_cuda = True) >> > mask = srng.binomial (n=1, p=1-dropout_rate,size=params.shape,dtype = >> > floatX) >> > >> > otuput = params *mask. >> > >> > doing this operation is doing two things. >> > >> > 1. If I use the libgpuarray backend, and I have two GPUS, the mask is >> being >> > created in another GPU from the one I have used in config.device. >> > 2. If I use it with use_cuda = False, it is creating the data in the >> same >> > GPU, but then the code does not run on GPU.. >> > >> > Appreciate any help. >> > >> > -- >> > >> > --- >> > 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. >> >> >> -- >> Pascal >> > -- --- 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.
