assign() got an unexpected keyword argument 'validate_shape'

[bersbersbers]

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow):

import tensorflow as tf

a = tf.Variable(2)
a.assign(5)
assert a.numpy() == 5

# ValueError: Shapes () and (2,) are incompatible
a.assign([1,2])  

# TypeError: assign() got an unexpected keyword argument 'validate_shape'
a.assign([1,2], validate_shape=False)

# ValueError: Shapes () and (2,) are incompatible
tf.compat.v1.assign(a, [1,2], validate_shape=False)  

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Windows 10, Linux
• TensorFlow installed from (source or binary): pip
• TensorFlow version (use command below): 2.0.0, 2.1.0
• Python version: 3.7

Describe the current behavior

tf.assign had a validate_shape parameter that Variable.assign seems to be missing.

In addition, the docs say:

If you want to change the shape of a variable later you have to use an assign Op with validate_shape=False.

https://www.tensorflow.org/api_docs/python/tf/Variable

How should one change the shape of a variable?

Code to reproduce the issue
See above.

[bersbersbers]

According to #30840, one now has to use a = tf.Variable(2, Shape=tf.TensorShape(None)) (compare #26104 and f9f2547).

Still, this looks like a documentation issue:

you have to use an assign Op with validate_shape=False.

is incorrect and should be replaced by

this is only possible if the variable has been initiliazed with shape=TensorShape(None) .

[rushabh-v]

Still, this looks like a documentation issue:

you have to use an assign Op with validate_shape=False.

is incorrect and should be replaced by

this is only possible if the variable has been initiliazed with shape=TensorShape(None) .

@MarkDaoust Can I open a PR as my first contribution that makes this change?