1 comment

[ 2.5 ms ] story [ 24.5 ms ] thread
While the post by Joel Grus was intended as a joke. I thought it would be interesting to use FizzBuzz as an example to illustrate how TensorFlow represents computations as graphs. This <20 lines version relies tf.mod / tf.argmin / tf.concat to deterministically compute results.

http://joelgrus.com/2016/05/23/fizz-buzz-in-tensorflow/