SentimentAnalysisModelingKerasWithCNTKBackend failed with error mkl err (-127)

[stluislu]

Environment:
Windows 10 x64
Python 3.5.2
Keras 2.1.1
CNTK CPU-Only 2.3

Outputs:
PS F:\Study\LuisSentimentAnalysis> az ml experiment submit -c local SentimentExtraction.py
RunId: LuisSentimentAnalysis_1511486594580

Executing user inputs .....

Using CNTK backend
Selected CPU as the process wide default device.
F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\keras\backend\cntk_backend.py:18: 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. '
Building model...
Train on 105 samples, validate on 106 samples
Epoch 1/2
Traceback (most recent call last):
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\azureml\logging_exec_wrapper.py", line 159, in execute
raise exc
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\azureml\logging_exec_wrapper.py", line 157, in execute
run_name="main")
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\runpy.py", line 254, in run_path
pkg_name=pkg_name, script_name=fname)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\runpy.py", line 96, in _run_module_code
mod_name, mod_spec, pkg_name, script_name)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "SentimentExtraction.py", line 179, in
model = train_model()
File "SentimentExtraction.py", line 99, in train_model
validation_data=(x_test, y_test))
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\keras\models.py", line 960, in fit
validation_steps=validation_steps)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\keras\engine\training.py", line 1650, in fit
validation_steps=validation_steps)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\keras\engine\training.py", line 1213, in _fit_loop
outs = f(ins_batch)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\keras\backend\cntk_backend.py", line 1815, in call
input_dict, self.trainer_output)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\cntk\train\trainer.py", line 171, in train_minibatch
output_map, device)
File "F:\Users\stlui\AppData\local\AmlWorkbench\Python\lib\site-packages\cntk\cntk_py.py", line 2853, in train_minibatch
return _cntk_py.Trainer_train_minibatch(self, *args)
RuntimeError: mkl err (-127)

[CALL STACK]
> Microsoft::MSR::CNTK::ConvolutionEngine:: MaxUnpooling
- Microsoft::MSR::CNTK::ConvolutionEngine:: Forward (x2)
- CNTK::Dictionary:: ~Dictionary (x2)
- CNTK::Internal:: UseSparseGradientAggregationInDataParallelSGD
- CNTK::Dictionary:: ~Dictionary
- CNTK::Internal:: UseSparseGradientAggregationInDataParallelSGD
- CNTK::Function:: Forward
- CNTK:: CreateTrainer
- CNTK::Trainer:: TotalNumberOfUnitsSeen
- CNTK::Trainer:: TrainMinibatch
- PyInit__cntk_py (x2)
- PyCFunction_Call
- PyEval_GetFuncDesc

Execution Details

RunId: LuisSentimentAnalysis_1511486594580