Using rounding when converting constants in BF16 propagation.
authorYuanzhong Xu <yuanzx@google.com>
Fri, 23 Mar 2018 18:41:05 +0000 (11:41 -0700)
committerTensorFlower Gardener <gardener@tensorflow.org>
Sun, 25 Mar 2018 09:47:47 +0000 (02:47 -0700)
commit1ece83d0cc98d09447e9a142605934328e18d15a
treeb48f57c5aee78a659615e3229423b4efb0808dd6
parent99f39f590128ca1d5e1d1645c69a147938e481bc
Using rounding when converting constants in BF16 propagation.

PiperOrigin-RevId: 190251081
tensorflow/compiler/xla/literal_util.cc
tensorflow/compiler/xla/literal_util.h
tensorflow/compiler/xla/service/bfloat16_propagation.cc