# Test tf.gather_nd behavior on GPU
# Usage: tf_gather_test.py <side> <margin>
python tf_gather_test.py 5 3This script first generates an input tensor of shape side x side, with
values 1 through side * side, in row-major order. This is useful for
tagging which elements of input are being gathered.
Secondly, it generates an indices tensor that contains all of the indices in
the input, plus some additional out-of-bounds indices. So, for example, it
generates the cartesian product of indices [-2, -1, 0, 1, 2, 3, 4, 5], for
side=4, margin=2
Finally, it prints out the results of the gather_nd call. gather_nd is
supposed to output 0 for any out-of-bounds indices. In the case of Python
tf, this is true.
For the call with side=5, margin=3, the output is:
[[ 0 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 1 2 3 4 5 0 0 0]
[ 0 0 0 6 7 8 9 10 0 0 0]
[ 0 0 0 11 12 13 14 15 0 0 0]
[ 0 0 0 16 17 18 19 20 0 0 0]
[ 0 0 0 21 22 23 24 25 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0 0]
[ 0 0 0 0 0 0 0 0 0 0 0]]
git clone https://github.com/hrbigelow/tfjs-tests.git
cd tfjs-tests
npm update
npm run devThis web app does the same thing as tf_gather_test.py but using
Tensorflow.js. In this case, tf.gatherND retrieves values from the input
even for out-of-bounds indices. They are reported to the console. The pattern
of retrieval for side=5, margin=3 is:
Tensor
[[1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ],
[1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 ],
[1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 2 , 3 ],
[1 , 1 , 1 , 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 ],
[3 , 4 , 5 , 6 , 7 , 8 , 9 , 10, 11, 12, 13],
[8 , 9 , 10, 11, 12, 13, 14, 15, 16, 17, 18],
[13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],
[18, 19, 20, 21, 22, 23, 24, 25, 25, 25, 25],
[23, 24, 25, 25, 25, 25, 25, 25, 25, 25, 25],
[25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25],
[25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25]]