Fix incorrect reference of np.assert_allclose (should be np.testing.assert_allclose)

In the docstring of tf.debugging.assert_near, the numpy compatibility
part incorrectly uses np.assert_allclose.

This should be np.testing.assert_allclose instead.

This PR fixes the incorrect docstring.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
This commit is contained in:
Yong Tang 2020-05-23 23:02:55 +00:00
parent ae76544efc
commit b9f941a53f
1 changed files with 6 additions and 6 deletions

View File

@ -750,9 +750,9 @@ def assert_near_v2(x, y, rtol=None, atol=None, message=None, summarize=None,
statically known. statically known.
@compatibility(numpy) @compatibility(numpy)
Similar to `numpy.assert_allclose`, except tolerance depends on data type. Similar to `numpy.testing.assert_allclose`, except tolerance depends on data
This is due to the fact that `TensorFlow` is often used with `32bit`, `64bit`, type. This is due to the fact that `TensorFlow` is often used with `32bit`,
and even `16bit` data. `64bit`, and even `16bit` data.
@end_compatibility @end_compatibility
""" """
return assert_near(x=x, y=y, rtol=rtol, atol=atol, summarize=summarize, return assert_near(x=x, y=y, rtol=rtol, atol=atol, summarize=summarize,
@ -802,9 +802,9 @@ def assert_near(
Op that raises `InvalidArgumentError` if `x` and `y` are not close enough. Op that raises `InvalidArgumentError` if `x` and `y` are not close enough.
@compatibility(numpy) @compatibility(numpy)
Similar to `numpy.assert_allclose`, except tolerance depends on data type. Similar to `numpy.testing.assert_allclose`, except tolerance depends on data
This is due to the fact that `TensorFlow` is often used with `32bit`, `64bit`, type. This is due to the fact that `TensorFlow` is often used with `32bit`,
and even `16bit` data. `64bit`, and even `16bit` data.
@end_compatibility @end_compatibility
""" """
message = message or '' message = message or ''