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

shorten tf.shape docstring

Browse Source 
clarify when it's different from `x.shape`
This commit is contained in:
Janosh Riebesell 2020-05-20 06:56:24 +02:00 committed by GitHub
parent f3f5afbe7b
commit 749bd23af6
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
tensorflow/python/ops/array_ops.py
View File

@ -535,19 +535,16 @@ def shape_v2(input, out_type=dtypes.int32, name=None):
# pylint: disable=redefined-builtin
"""Returns the shape of a tensor.
This operation returns a 1-D integer tensor representing the shape of `input`.
This represents the minimal set of known information at definition time.
`tf.shape` returns a 1-D integer tensor representing the shape of `input`.
For example:
>>> t = tf.constant([[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]])
>>> tf.shape(t)
<tf.Tensor: shape=(3,), dtype=int32, numpy=array([2, 2, 3], dtype=int32)>
>>> tf.shape(t).numpy()
array([2, 2, 3], dtype=int32)
Note: When using symbolic tensors, such as when using the Keras functional
API, tf.shape() will return the shape of the symbolic tensor.
Note: When using symbolic tensors, such as when using the Keras API,
tf.shape() will return the shape of the symbolic tensor.
>>> a = tf.keras.layers.Input((None, 10))
>>> tf.shape(a)
@ -558,17 +555,12 @@ def shape_v2(input, out_type=dtypes.int32, name=None):
>>> a.shape
TensorShape([None, None, 10])
However, when defining custom layers and models that will be run in graph mode
at some point, prefer `tf.shape(x)` over `x.shape`. `x.shape` is the static shape
of `x` and usually evaluates to `None` in the first dimension during graph
construction (to represent the as yet unknown batch size). This can cause problems in
function calls like `tf.zeros(x.shape[0])` which don't support `None` values.
`tf.shape(x)` on the other hand gives the dynamic shape of `x` which isn't
evaluated until training/predicting begins where the full shape of `x` is known.
(The first `None` represents the as yet unknown batch size.)
`tf.shape` and `Tensor.shape` should be identical in eager mode. Within
`tf.function` or within a `compat.v1` context, not all dimensions may be
known until execution time.
known until execution time. Hence when defining custom layers and models
for graph mode, prefer the dynamic `tf.shape(x)` over the static `x.shape`.
Args:
input: A `Tensor` or `SparseTensor`.