You can ignore it or update Theano to the dev version:

http://www.deeplearning.net/software/theano/install.html#bleeding-edge-install-instructions

they are fixed in the master of Theano.

On Thu, Aug 25, 2016 at 9:31 PM, bwang <[email protected]> wrote:

> Hi,
>
> I am a fresher in theano and linux (Ubuntu 14.04.3 LTS). I installed the
> theano based on anaconda this week (PYTHON 3.5), and my .theanorc file is
>
> [global]
> floatX=float32
> device=cpu
> [cuda]
> root=/usr/lib/nvidia-cuda-toolkit
> [nvcc]
> flags=-D_FORCE_INLINES
>
> I tested it via theano.test() today, and got some error messages
> (repeatedly) so far (theano.test() is still running):
>
>
>
> ERROR (theano.gof.opt): Optimization failure due to: local_softmax_dnn_grad
>
> ERROR (theano.gof.opt): TRACEBACK:
>
> ERROR (theano.gof.opt): Traceback (most recent call last):
>
> File 
> "/home/norman_lab/anaconda3/lib/python3.5/site-packages/theano/gof/opt.py",
> line 1492, in process_node
>
> replacements = lopt.transform(node)
>
> File 
> "/home/norman_lab/anaconda3/lib/python3.5/site-packages/theano/sandbox/cuda/dnn.py",
> line 1696, in local_softmax_dnn_grad
>
> and (isinstance(node.inputs[0].owner.op, HostFromGpu)
>
> AttributeError: 'NoneType' object has no attribute 'op'
>
>
>
> BUG IN FGRAPH.REPLACE OR A LISTENER !!>> <class 'TypeError'> ('The type of
> the replacement must be compatible with the type of the original
> Variable.', Sum{axis=[0], acc_dtype=float64}.0, HostFromGpu.0,
> TensorType(int16, vector), TensorType(float32, vector),
> 'local_gpu_careduce') local_gpu_careduce
>
> ERROR (theano.gof.opt): Optimization failure due to: local_gpu_careduce
>
> ERROR (theano.gof.opt): TRACEBACK:
>
> ERROR (theano.gof.opt): Traceback (most recent call last):
>
> File 
> "/home/norman_lab/anaconda3/lib/python3.5/site-packages/theano/gof/opt.py",
> line 1526, in process_node
>
> fgraph.replace_all_validate(repl_pairs, reason=lopt)
>
> File 
> "/home/norman_lab/anaconda3/lib/python3.5/site-packages/theano/gof/toolbox.py",
> line 259, in replace_all_validate
>
> fgraph.replace(r, new_r, reason=reason, verbose=False)
>
> File 
> "/home/norman_lab/anaconda3/lib/python3.5/site-packages/theano/gof/fg.py",
> line 474, in replace
>
> str(reason))
>
> TypeError: ('The type of the replacement must be compatible with the type
> of the original Variable.', Sum{axis=[0], acc_dtype=float64}.0,
> HostFromGpu.0, TensorType(int16, vector), TensorType(float32, vector),
> 'local_gpu_careduce')
>
>
>
> I did a lot of search but still cannot find the solution. Can any one help
> me out?
>
>
> Many thanks!
>
>
> [global]
> floatX=float32
> device=cpu
> [cuda]
> root=/usr/lib/nvidia-cuda-toolkit
> [nvcc]
> flags=-D_FORCE_INLINES
>
> --
>
> ---
> 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