I'm not able to help with the information provided. I will need at least
the full error message. If you also a script that reproduce that, it could
also help.

Fred

On Wed, Apr 19, 2017 at 3:15 AM Arthi Venkataraman <[email protected]>
wrote:

> Hi,
> Getting TypeError: iteration over a 0-d array  with Theano version
> '0.9.0.dev-388805f946685e86225cdf602eb8a4f0059f9667'
>
> Happens when calling the defined function.
> self.trainfunc =
> theano.function([inquery,inmemory,learningrate,yin,yin1,yin2,yin4],
> cost,updates=updates,allow_input_downcast = True)
>
> cost,v1,v2,v3 = mem_network.trainfunc(query,data_train,
> learningrate,yval1,yval,yval2,yval)
>
> Error does not come when running the same in debug mode as called below :
> mode =
> theano.compile.MonitorMode(post_func=detect_nan).excluding('local_elemwise_fusion',
> 'inplace')
> self.trainfunc =
> theano.function([inquery,inmemory,learningrate,yin,yin1,yin2,yin4],
> [cost,v1,v2,output],updates=updates,allow_input_downcast = True, mode=mode)
>
> There are no zero dimensional arrays in the call.
> These are the values of the parameters.
> query = np.zeros((sent_len),dtype=np.int)
>  yval1 = numpy.zeros((batch_size,sent_len,embsize),dtype=numpy.float64)  #
> Batchsize * Seqlen * Emb dim  3d
>   yval = numpy.zeros((batch_size,sent_len,embsize),dtype=numpy.float64)  #
> Batchsize * Seqlen * Emb dim 3d
>   yval2 = numpy.zeros((sent_len,vocabsize),dtype=numpy.float64)  # Seqlen
> * Vocabsize 2d
>
>
>
>
> --
>
> ---
> 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.

Reply via email to