diff options
author | 2017-10-02 16:27:21 -0700 | |
---|---|---|
committer | 2017-10-02 16:30:57 -0700 | |
commit | 9b027db459ff771c246a266ac3ec40cfbb4a63ce (patch) | |
tree | 12c51bd990531676e695719739107257e0d3ab4e /tensorflow/core/kernels/where_op_gpu_impl_5.cu.cc | |
parent | 6ce50b2a991270eeed620050eff29b5b91422a8e (diff) |
[tf.data] Use the user-provided type when converting to a NumPy array.
This eases Windows-vs-Linux `np.int32`-vs-`np.int64` issues when no types
are given. Fixes #13431.
PiperOrigin-RevId: 170772767
Diffstat (limited to 'tensorflow/core/kernels/where_op_gpu_impl_5.cu.cc')
0 files changed, 0 insertions, 0 deletions