Merge pull request #17544 from DavidNorman/add-half-constraints
authorPeter Hawkins <hawkinsp@cs.stanford.edu>
Tue, 3 Apr 2018 12:20:51 +0000 (08:20 -0400)
committerGitHub <noreply@github.com>
Tue, 3 Apr 2018 12:20:51 +0000 (08:20 -0400)
Register half in some ops which support all floating point types

1  2 
tensorflow/core/ops/nn_ops.cc

Simple merge