experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Fix numpy-ndarray bug

Browse Source 
On zSystems, the ndarray is not retaining the values for nparray. Defining the data type is required to get the desired behavior.
This commit is contained in:
Shahid 2017-09-11 16:52:13 +05:30 committed by GitHub
parent 702d595822
commit 76ad78c6dd
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tensorflow/python/framework/tensor_util.py
View File

@ -627,7 +627,7 @@ def _ConstantValue(tensor, partial):
elif tensor.op.type == "Rank": elif tensor.op.type == "Rank":
input_shape = tensor.op.inputs[0].get_shape() input_shape = tensor.op.inputs[0].get_shape()
if input_shape.ndims is not None: if input_shape.ndims is not None:
return np.ndarray(shape=(), buffer=np.array([input_shape.ndims]), return np.ndarray(shape=(), buffer=np.array([input_shape.ndims], dtype=np.int32),
dtype=np.int32) dtype=np.int32)
else: else:
return None return None