PRelu - 6 vs 7

Files changed (1) hide show
  1. PRelu6 → PRelu7 +4 -3
PRelu6 → PRelu7 RENAMED
@@ -1 +1 @@
1
1
  PRelu takes input data (Tensor<T>) and slope tensor as input, and produces one
2
2
  output data (Tensor<T>) where the function f(x) = slope * x for x < 0,
3
3
  f(x) = x for x >= 0., is applied to the data tensor elementwise.
4
+ This operator supports **unidirectional broadcasting** (tensor slope should be unidirectional broadcastable to input tensor X); for more details please check Broadcasting in ONNX <https://github.com/onnx/onnx/blob/master/docs/Broadcasting.md>_.
4
5
  **Inputs**
5
6
  * **X** (heterogeneous) - **T**:
6
7
  Input tensor
7
8
  * **slope** (heterogeneous) - **T**:
8
- Slope tensor. If Slope is of size 1, the value is sharedacross
9
+ Slope tensor. The shape of slope can be smaller then first input X;
9
- different channels
10
+ if so, its shape must be unidirectional broadcastable to X
10
11
  **Outputs**
11
12
  * **Y** (heterogeneous) - **T**:
12
- Output tensor
13
+ Output tensor (same size as X)
13
14
  **Type Constraints**
14
15
  * **T** in (
15
16
  tensor(double),
16
17
  tensor(float),
17
18
  tensor(float16)
18
19
  ):
19
20
  Constrain input and output types to float tensors.