undefined | Better HN

0 pointscglace10y ago0 comments

Using all inputs and 6 layers of varying sizes. After about 500 iterations. http://i.imgur.com/x1MOpvl.jpg

0 comments

1 comments · 1 top-level

Just 100 iterations, learning rate 0.03, activation tanh, regularization L2, rate 0.01. The network is 8,8,8 neurons per layer.

j / k navigate · click thread line to collapse