tf.math.reciprocal lacks support of int input on CPU, inconsistent with the documentation about tensorflow HOT 5 CLOSED

@sushreebarsa Sure, thanks for your effort in fixing it.

from tensorflow.

@drewshark Before applying tf.math.reciprocal, could you try to explicitly cast your integer input tensor to a supported floating-point data type using tf.cast as follows;

import os
os.environ['CUDA_VISIBLE_DEVICES'] = ''
import numpy as np
import tensorflow as tf
x = tf.constant(np.random.randint(-50, 50, ()), dtype='int64')
x = tf.cast(x, dtype=tf.float64)
tf.math.reciprocal(x)

It might resolve the issue. Please find the gist here for reference.
Thank you!

from tensorflow.

This issue is stale because it has been open for 7 days with no activity. It will be closed if no further activity occurs. Thank you.

from tensorflow.

@drewshark Glad it helped you!
Thank you for your response here.
Closing the issue for now.

from tensorflow.

Are you satisfied with the resolution of your issue?
Yes
No

from tensorflow.