Hello,

I want to use GPUs with Theano, and read here 
(http://deeplearning.net/software/theano/tutorial/examples.html#using-random-numbers)
 
that I have to import this

from theano.sandbox.rng_mrg import MRG_RandomStreams as RandomStreams

instead of this

from theano.tensor.shared_randomstreams import RandomStreams

However, I have the feeling that both are working with GPU, since I got 
random numbers as expected when running the above code with the GPU :
import theano
# from theano.sandbox.rng_mrg import MRG_RandomStreams as RandomStreams
from theano.tensor.shared_randomstreams import RandomStreams


if __name__ == '__main__':
    rng = RandomStreams()

    def get_random():
        out = rng.uniform((2,2), dtype=theano.config.floatX)
        fonction = theano.function(inputs=[],
                                   outputs=[out],
                                   name='test'
                                   )
        return fonction

    fonc = get_random()
    for i in range(5):
        print(fonc())

Is it normal ? I have the confirmation that i'm running the code on the GPU,
Using gpu device 0: GeForce GTX TITAN X (CNMeM is enabled with initial size: 
90.0% of memory, cuDNN not available)

So why is it working ?

-- 

--- 
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.

Reply via email to