Initialization function for the left part intercept. Default is Zeros. You can also pass in corresponding string representations such as 'zero' or 'normal', etc. for simple init methods in the factory method.
Initialization function for the left part slope. Default is Xavier. You can also pass in corresponding string representations such as 'glorot_uniform', etc. for simple init methods in the factory method.
Initialization function for the right part intercept. Default is Xavier. You can also pass in corresponding string representations such as 'glorot_uniform', etc. for simple init methods in the factory method.
Initialization function for the right part slope. Default is Ones. You can also pass in corresponding string representations such as 'one' or 'normal', etc. for simple init methods in the factory method.
Array of Int. The axes along which to share learnable parameters for the activation function. Default is null. For example, if the incoming feature maps are from a 2D convolution with output shape (batch, height, width, channels), and you wish to share parameters across space so that each filter only has one set of parameters, set 'sharedAxes = Array(1,2)'.
A Single Shape, does not include the batch dimension.
Initialization function for the left part slope.
Initialization function for the left part slope. Default is Xavier. You can also pass in corresponding string representations such as 'glorot_uniform', etc. for simple init methods in the factory method.
Initialization function for the right part slope.
Initialization function for the right part slope. Default is Ones. You can also pass in corresponding string representations such as 'one' or 'normal', etc. for simple init methods in the factory method.
Build graph: some other modules point to current module
Build graph: some other modules point to current module
upstream variables
Variable containing current module
A Single Shape, does not include the batch dimension.
A Single Shape, does not include the batch dimension.
Array of Int.
Array of Int. The axes along which to share learnable parameters for the activation function. Default is null. For example, if the incoming feature maps are from a 2D convolution with output shape (batch, height, width, channels), and you wish to share parameters across space so that each filter only has one set of parameters, set 'sharedAxes = Array(1,2)'.
Initialization function for the left part intercept.
Initialization function for the left part intercept. Default is Zeros. You can also pass in corresponding string representations such as 'zero' or 'normal', etc. for simple init methods in the factory method.
Initialization function for the right part intercept.
Initialization function for the right part intercept. Default is Xavier. You can also pass in corresponding string representations such as 'glorot_uniform', etc. for simple init methods in the factory method.
(Since version 0.3.0) please use recommended saveModule(path, overWrite)
S-shaped Rectified Linear Unit. It follows: f(x) = tr + ar(x - tr) for x >= tr, f(x) = x for tr > x > tl, f(x) = tl + al(x - tl) for x <= tl.
When you use this layer as the first layer of a model, you need to provide the argument inputShape (a Single Shape, does not include the batch dimension).
Numeric type of parameter(e.g. weight, bias). Only support float/double now.