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

GroupNormalization core op (#17781)

Browse Source
This commit is contained in:
Tomasz Dołbniak
2023-06-01 08:49:02 +02:00
committed by GitHub
parent 0b6b16c83a
commit 02124aece4
21 changed files with 713 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/ops/normalization/GroupNormalization_12.md
View File

@@ -38,9 +38,9 @@ The operation is applied per batch, per group of channels. This means that the e
* **1**: ``data`` - The input tensor to be normalized. The type of this tensor is *T*. The tensor's shape is arbitrary but the first two dimensions are interpreted as ``batch`` and ``channels`` respectively. **Required.**
* **2**: ``scale`` - 1D tensor of type *T* containing the scale values for each group. The expected shape of this tensor is ``[C]`` where ``C`` is the number of channels in the ``data`` tensor. **Required.**
* **2**: ``scale`` - 1D tensor of type *T* containing the scale values for each channel. The expected shape of this tensor is ``[C]`` where ``C`` is the number of channels in the ``data`` tensor. **Required.**
* **3**: ``bias`` - 1D tensor of type *T* containing the bias values for each group. The expected shape of this tensor is ``[C]`` where ``C`` is the number of channels in the ``data`` tensor. **Required.**
* **3**: ``bias`` - 1D tensor of type *T* containing the bias values for each channel. The expected shape of this tensor is ``[C]`` where ``C`` is the number of channels in the ``data`` tensor. **Required.**
**Outputs**