building an array in tensorflow

Question

I have:

arr_1 = tf.random_uniform([3],0.1,0.2)
arr_2 = tf.random_uniform([3],-0.1,0.1)

I am trying to create an arr_3 which has a size of 6 and alternates elements from arr_1 and arr_2. How should I go about this? Since tensors are read-only as such, I don't think I can index them and assign values the same way I would with numpy. Any solutions?

Why not first do that using numpy and convert arr_3 into tensor? — tourist
– tourist, Commented Jun 19, 2018 at 18:06
I am doing this inside a map function for an object of the Dataset API. If I use numpy functions, it does not generate a random number of every single call, but instead uses the number it generated in the first call. — TheBlueNotebook
– TheBlueNotebook, Commented Jun 19, 2018 at 18:14

Abhishek Mishra · Accepted Answer · 2018-06-19 18:33:59Z

2

This might work. The idea is based on stacking the data and then flattening it.

arr_3 = tf.reshape(tf.stack([arr_1, arr_2], axis=1), [-1]))

answered Jun 19, 2018 at 18:33

Abhishek Mishra

2,0041 gold badge19 silver badges14 bronze badges

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

Collectives™ on Stack Overflow

building an array in tensorflow

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related