TensorFlow Variables and tf.assign() with string

Question

I am new to tensorflow , here is a quick question , this is my code

session=tf.Session()
x=tf.Variable(str)
valueOfX=session.run(x.assign('xyz'))
print(valueOfX)

Why the output is =>

b'xyz'

But when I use int as datatype and assign an integer , the assignment is fine .

By "the assignment is fine" what do you mean? You get the expected output? Can you show us that please. — DarkCygnus
– DarkCygnus, Commented Jun 26, 2017 at 17:23

mrry · Accepted Answer · 2017-06-26 17:23:41Z

2

This confusion arises because Python 3 uses a Unicode string representation for string literals. The printed representation b'xyz' means that valueOfX is a bytes object. TensorFlow uses bytes as the internal representation of all string tensors and variables, and (when using Python 3) implicitly converts str literals, such as the 'xyz' in your code, to bytes using a UTF-8 unicode encoding.

answered Jun 26, 2017 at 17:23

mrry

126k27 gold badges404 silver badges401 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Daniyal Javaid Over a year ago

so my output is completely fine ? or there is something which can be done to avoid this 'b' ?

mrry Over a year ago

Yes, it is fine. Assuming that your strings use only ASCII characters, you can disregard the b prefix. If you use non-ASCII Unicode characters in your program, you may want to print valueOfX.decode() instead (and specify the appropriate encoding if you're not using UTF-8).

Daniyal Javaid Over a year ago

thanks, i tried the decoding statement and it works !

Collectives™ on Stack Overflow

TensorFlow Variables and tf.assign() with string

1 Answer 1

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related