Can't run verification.ipynb

`---------------------------------------------------------------------------
MXNetError Traceback (most recent call last)
in ()
45 for i in range(len(ver_list)):
46 # Compute and print validation accuracies
---> 47 acc2, std2, xnorm, embeddings_list = test(ver_list[i], model, batch_size, nfolds)
48 print('[%s]XNorm: %f' % (ver_name_list[i], xnorm))
49 print('[%s]Accuracy-Flip: %1.5f+-%1.5f' % (ver_name_list[i], acc2, std2))

2 frames
in test(data_set, mx_model, batch_size, nfolds, data_extra, label_shape)
27 model.forward(db, is_train=False)
28 net_out = model.get_outputs()
---> 29 _embeddings = net_out[0].asnumpy()
30 time_now = datetime.datetime.now()
31 diff = time_now - time0

/usr/local/lib/python3.6/dist-packages/mxnet/ndarray/ndarray.py in asnumpy(self)
1978 self.handle,
1979 data.ctypes.data_as(ctypes.c_void_p),
-> 1980 ctypes.c_size_t(data.size)))
1981 return data
1982

/usr/local/lib/python3.6/dist-packages/mxnet/base.py in check_call(ret)
250 """
251 if ret != 0:
--> 252 raise MXNetError(py_str(_LIB.MXGetLastError()))
253
254

MXNetError: [06:40:47] src/operator/./leaky_relu-inl.h:311: Check failed: s == -1 Cannot broadcast gamma to data. gamma: [1,64,1,1], data: [1,64,112,112]`

Running the correct config as per the instructions apart from changing mxnet version to match cuda version:

!pip install mxnet-cu100

Using data faces_ms1m_112x112.zip using the pretrained LResNet100E-IR.

Start with verification.ipynb and reproduce the failure using faces_ms1m_112x112.zip, the pretrained LResNet100E-IR model, and the documented mxnet-cu100 installation. Inspect the test call that reaches net_out[0].asnumpy() and the reported leaky_relu broadcast error. Done means the notebook completes verification without this MXNetError under the stated CUDA-compatible setup.