A hardware-accelerated machine intelligence library for the web (pair-code.github.io) 59 points by obulpathi 8y ago ↗ HN
[–] gradys 8y ago ↗ Here's the associated Google Research Blog post:https://research.googleblog.com/2017/08/harness-power-of-mac... [–] dgacmu 8y ago ↗ Tl;dr Tensorflow and numpy-like API. Differs from tensorfire in that it can also do backprop (training)
[–] dgacmu 8y ago ↗ Tl;dr Tensorflow and numpy-like API. Differs from tensorfire in that it can also do backprop (training)
[–] mncharity 8y ago ↗ http://playground.tensorflow.org/ is fun and worth looking at.As I recall, Google created it to give their engineers an introductory feel for DL, and then open sourced it.
[–] xamlhacker 8y ago ↗ "Currently our demos do not support Mobile, Firefox, and Safari. Please view them on desktop Chrome for now." That's a bummer. [–] nsthorat 8y ago ↗ This will be fixed in ~2 weeks, max.
[–] mamp 8y ago ↗ It looks like this is the easiest way to train neural networks with GPU acceleration on my Radeon Pro (i.e. non CUDA)...Thank you JavaScript!
7 comments
[ 3.2 ms ] story [ 27.8 ms ] threadhttps://research.googleblog.com/2017/08/harness-power-of-mac...
As I recall, Google created it to give their engineers an introductory feel for DL, and then open sourced it.
Thank you JavaScript!