Xavier 2#
The Xavier 2 initializer draws from a uniform distribution [-limit, limit] where limit is equal to (6 / (fanIn + fanOut)) ** 0.25. This initializer is best suited for layers that feed into an activation layer that outputs values between -1 and 1 such as Hyperbolic Tangent and Softsign.
Parameters#
This initializer does not have any parameters.
Example#
use Rubix\ML\NeuralNet\Initializers\Xavier2;
$initializer = new Xavier2();
References#
-
X. Glorot et al. (2010). Understanding the Difficulty of Training Deep Feedforward Neural Networks. ↩
Last update:
2021-03-03