IntenseWebs/openvino
3
0 Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[Spec] Prelu - Remove 1D slope restriction and update broadcast info (#5951)

Browse Source 
* Remove 1D restriction and update broadcast info

* Add link to Broadcast Rules description
This commit is contained in:
Katarzyna Mitrus authored and GitHub committed 2021-06-02 14:30:15 +02:00
1 parent 3c9092ffb8
commit 5f54ec559a
1 file changed
+5 -5
Download Patch File Download Diff File Expand all files Collapse all files
docs/ops/activation/PReLU_1.md
+5 -5
View File
@@ -19,7 +19,7 @@ PReLU(x) = \left\{\begin{array}{r}
\end{array}\right.
\f]
where α is a learnable parameter and corresponds to the negative slope, per channel, defined by the second input `slope`.
where α is a learnable parameter and corresponds to the negative slope, defined by the second input `slope`.
Another mathematical representation that may be found in other references:
@@ -32,13 +32,13 @@ PReLU(x) = \max(0, x) + \alpha\cdot\min(0, x)
**Inputs**
* **1**: `data`. A tensor of type `T` and arbitrary shape. **Required**.
* **2**: `slope`. 1D tensor of type `T`. Tensor with negative slope values, one per channel dimension of `data` input tensor. **Required**.
* **Note**: Channels dimension corresponds to second dimension of `data` input tensor. If `data` rank is less than 2, the number of channels is 1.
* **1**: `data`. A tensor of type *T* and arbitrary shape. **Required**.
* **2**: `slope`. A tensor of type *T* and rank greater or equal to 1. Tensor with negative slope values. **Required**.
* **Note**: Channels dimension corresponds to the second dimension of `data` input tensor. If `slope` input rank is 1 and its dimension is equal to the second dimension of `data` input, then per channel broadcast is applied. Otherwise `slope` input is broadcasted with numpy rules, description is available in [Broadcast Rules For Elementwise Operations](../broadcast_rules.md).
**Outputs**
* **1**: The result of element-wise *PReLU* operation applied to `data` input tensor with negative slope values from `slope` input tensor. A tensor of type `T` and the same shape as `data` input tensor.
* **1**: The result of element-wise *PReLU* operation applied to `data` input tensor with negative slope values from `slope` input tensor. A tensor of type *T* and the same shape as `data` input tensor.
**Types**