I have one tensor "x" with dimension [n, 6, 6, 20] and another tensor "seed" with dimension [6, 6, 6]. I want the following operation:
x[ 1: , : , : , :6 ] = seed, where sum of a row of seed is not zero (By row I mean the last dimension of seed).
I have tried turning them into numpy arrays but part of my code requires eager execution to be disabled, which means I can not do that without changing most of the code. I've also tried using masks for the seed tensor, to no avail. I'm using Tensorflow 2.7.0 and Python 3.8.7
Is there no easy way of doing this?
Also if there is something wrong with my question, please educate me.