ML::Autograd::SigmoidBackward
Inherits ML::Autograd::GradFn / Reference / Object
Gradient for Sigmoid: out = 1 / (1 + exp(-x)) d(sigmoid)/dx = out * (1 - out)
Constructors
new(output_data : Tensor)
SourceInherits ML::Autograd::GradFn / Reference / Object
Gradient for Sigmoid: out = 1 / (1 + exp(-x)) d(sigmoid)/dx = out * (1 - out)