CNTK as backend with Kegra throws error

[neocsbee]

Hi,

I am using CNTK as the backend for my Kegra install. (found here: https://github.com/tkipf/keras-gcn?files=1)
When I tried a sample run of Kegra, I get the following error: (\keras-gcn-master\kegra>python train.py)

Using CNTK backend
Selected CPU as the process wide default device.
Anaconda2\lib\site-packages\keras\backend\cntk_backend.py:21: UserWarning: CNTK backend warning: GPU is not detected. CNTK's CPU version is not fully optimized,please run with GPU to get better performance.
'CNTK backend warning: GPU is not detected. '
Loading cora dataset...
Dataset has 2708 nodes, 5429 edges, 1433 features.
Using local pooling filters...
Traceback (most recent call last):
File "train.py", line 57, in
Y = GraphConvolution(y.shape[1], support, activation='softmax')([H]+G)
File "\Anaconda2\lib\site-packages\keras\engine\topology.py", line 617, in call
output = self.call(inputs, **kwargs)
File "build\bdist.win-amd64\egg\kegra\layers\graph.py", line 71, in call
File "\Anaconda2\lib\site-packages\keras\backend\cntk_backend.py", line 523, in dot
return C.times(x, y)
File "\Anaconda2\lib\site-packages\cntk\internal\swig_helper.py", line 69, in wrapper
result = f(*args, **kwds)
File "\Anaconda2\lib\site-packages\cntk\ops_init_.py", line 1038, in times
return times(right, left, output_rank, infer_input_rank_to_map, name)

ValueError: Times: right operand 'Output('Select49_Output_0', [#], [* x 16])' shape '[* x 16]' dimension cannot be inferred from a left operand 'Input('input_1', [#], [])' shape '[]' free dimension.

[CALL STACK]
> CNTK::NDMask:: MaskedCount
- CNTK::NDMask:: MaskedCount
- CNTK::Function:: ~Function
- CNTK_ReleaseModel
- RtlRunOnceExecuteOnce
- InitOnceExecuteOnce
- _crtInitOnceExecuteOnce
- CNTK_ReleaseModel
- CNTK::Function:: RawOutputs
- CNTK::Internal:: UseSparseGradientAggregationInDataParallelSGD
- CNTK::Function:: ~Function
- CNTK_ReleaseModel
- RtlRunOnceExecuteOnce
- InitOnceExecuteOnce
- _crtInitOnceExecuteOnce
- CNTK_ReleaseModel

Thanks.