News for hackers. Stuff that matters.

A comment on Previewing Android P. Parent comment

Re:Previewing Android P by mcintyre1994 on Wednesday March 07, 2018 @05:46PM (#16540697)
It's in TensorFlow: https://www.tensorflow.org/api_docs/python/tf/nn/relu6

And even points to a paper: http://www.cs.utoronto.ca/~kriz/conv-cifar10-aug2010.pdf

I'm not sure why they didn't make it a variable maximum though, or what the advantage is over multiplying a sigmoid by 6.

[ Parent | on HN ]
Re:Previewing Android P by clarle on Wednesday March 07, 2018 @07:11PM (#16541175)
There's an explanation here:

https://stackoverflow.com/questions/47220595/why-the-6-in-re...

It's an arbitrary value, but 6 fits best into 8 bits.

[ Parent | on HN ]