| The
PRelu op takes input data tensor
| $X$, an input slope tensor $slope$,
| and produces one output tensor $Y$ of
| the same shape as $X.$ The op performs
| the element wise
PRelu operation,
| defined as
|
| $$y=prelu(x) =\begin{cases}slope
| * x & x < 0\x & otherwise\end{cases}$$
|
| Note, is slope is size 1, the value is
| shared across the channels, otherwise
| $X$ and $slope$ must be the same shape.
| See
Delving Deep into Rectifiers:
| Surpassing Human-Level Performance
| on
|
| ImageNet Classification
| for more information.
|
| Github Links:
|
| - https://github.com/pytorch/pytorch/blob/master/caffe2/operators/prelu_op.h
| - https://github.com/pytorch/pytorch/blob/master/caffe2/operators/prelu_op.cc
|