edzzn (2) [Avatar] Offline
#1
Hello.

I am currently in the chapter 3 of the book and for the most part I am liking it.

I have a small problem with the way we are defining the data types of constants, Variables and placeholders.

Why do you use "float" as the data type when Everyone else uses tf.float32?

(This code is at Listing 3.3 but I am seen it in all the code so far.

X = tf.placeholder("float")
Y = tf.placeholder("float")


I think you should change this before the book is published. There is no need to use "float" when every other tutorial, book or code in Tensorflow.org and most of the code in github use tf.float32 to define the variables.


I think this is just confusing. Or if there is another reason for using the "float" instead of tf.float32 please tell me.


Thanks,
Edisson Reinozo

Nishant Shukla (52) [Avatar] Offline
#2
Hi Edisson,

Totally agree! I'll fix that.