experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

make usage of is_training more clear (#9730)

Browse Source 
* make usage of is_training more clear

This is the continuation of #9723.

* Wrap lines
This commit is contained in:
Androbin 2017-05-06 21:07:06 +02:00 committed by Vijay Vasudevan
parent d5956e3e6a
commit 3358855e01
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
tensorflow/docs_src/community/style_guide.md
View File

@ -162,9 +162,9 @@ operation.
- `reuse`: `bool` indicator if the variable should be reused if - `reuse`: `bool` indicator if the variable should be reused if
it's present in the scope. it's present in the scope.
* Layers that behave differently during training should have: * Layers that behave differently during training should take:
- `is_training`: `bool` indicator if the graph is being built for training. - `is_training`: `bool` indicator to conditionally choose different
computation paths (e.g. using `tf.cond`) during execution.
Example: Example: