Crate caffe2op_hardsigmoid

Structs

GetHardSigmoidGradient
HardSigmoidFunctor
| Applies hard sigmoid operation to the | input data element-wise. | | The HardSigmoid operation takes one | input $X$, produces one output $Y$, | and is defined as: | | $$Y = max(0,min(1,x * alpha + beta))$$ | | Github Links: | | - https://github.com/pytorch/pytorch/blob/master/caffe2/operators/hard_sigmoid_op.h | | - https://github.com/pytorch/pytorch/blob/master/caffe2/operators/hard_sigmoid_op.cc |
HardSigmoidGradientFunctor
| HardSigmoidGradient takes both Y and | dY as well as an argument alpha and uses | this to update dX according to the chain | rule and derivatives of the hard sigmoid | function. |