How to convert array of Integers to binary encoding in TensorFlow or Numpy?

Question

I have an input of numbers, ranging on [0, 15], which I want to feed to my network as a binary code with 4 bits. For example, the input [0,1,4,7] should become [[0,0,0,0],[0,0,0,1],[0,1,0,0],[0,1,1,1]].

The tf.one_hot operation is close, but not exactly what I want. Is there any elegant way, either with Numpy or TensorFlow, to convert my input into its binary encoding, in order to feed it into my network?

My best solution was to use np.binary_repr for each value, and convert it from a string into an array of integers, but I feel this is not a good solution (converting twice, first into string, then into array).

Dan · Accepted Answer · 2018-03-08 15:02:00Z

1

Something like this:

np.unpackbits(np.array([[0,1,4,7]],np.uint8)).reshape(-1,4)[1::2,:]

I'm sure it can be refined but at least it's vectorized

Or this probably makes more sense:

np.unpackbits(np.array([[0,1,4,7]],np.uint8)).reshape(-1,8)[:,4:]

answered Mar 8, 2018 at 15:02

Dan

45.8k20 gold badges98 silver badges170 bronze badges

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

1 Comment

BlueMoon93 Over a year ago

The second option seems much cleaner. Good job

Collectives™ on Stack Overflow

How to convert array of Integers to binary encoding in TensorFlow or Numpy?

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related