Thank you for the code, but I was actually talking about the cudnn implementation. Are you aware of any example of that?
> Il giorno 06 gen 2017, alle ore 19:23, Ramana Subramanyam > <[email protected]> ha scritto: > > Hi, > I'm sorry about that. The command+k didn't work for linking the code. Here is > the one that I was referring to : https://github.com/lipiji/rnn-theano > > Ramana > >> On Thursday, January 5, 2017 at 5:39:37 PM UTC+1, Marco Ciccone wrote: >> It seems you didn't link any reference :) >> Btw I'm using this code: >> https://github.com/marcociccone/cudnn_rnn_theano_benchmarks >> >> you can try the bidirectional version using the argumetn -md 2 >> >> Thanks for your help >> >> Il giorno giovedì 5 gennaio 2017 01:07:44 UTC+1, Ramana Subramanyam ha >> scritto: >>> >>> Hi, >>> It would be better for us if you could share your code. For reference, you >>> can have a look at this repo for GRU and LSTM example. >>> >>> Ramana >>> >>>> On Monday, December 26, 2016 at 10:13:52 PM UTC+1, Marco Ciccone wrote: >>>> Hi, >>>> I'm trying to use the rnn operations recently added from cudnn. >>>> I've been able to use GRU and LSTM in the unidirectional mode, but when I >>>> try to use them also in both directions I come up with >>>> CUDNN_STATUS_BAD_PARAM exception. >>>> >>>> I think that the issue is about the memory allocated for the parameters >>>> and the hidden states. >>>> Is there an example or a test that I can use to compare my implementation? >>>> >>>> Thanks >>>> Marco > > -- > > --- > You received this message because you are subscribed to a topic in the Google > Groups "theano-users" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/theano-users/UdBy5KqOpyQ/unsubscribe. > To unsubscribe from this group and all its topics, 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.
