Update randomness.rst (#21337)

Bram Vanroy · facebook-github-bot · commit 38d68ad80330 · 2019-06-04T07:38:00.000-07:00
Summary: Following [this question on the forums](https://discuss.pytorch.org/t/reproducibility-and-performance/46504), I propose the following doc change. It clarifies that 'performance reduction' concerns the processing speed (and not the training accuracy). Related website commit: pytorch/pytorch.github.io#211 Pull Request resolved: #21337 Differential Revision: D15622151 Pulled By: soumith fbshipit-source-id: f0edeb20049f2ee715c400e7c57abb966864d621
diff --git a/docs/source/notes/randomness.rst b/docs/source/notes/randomness.rst
@@ -45,7 +45,7 @@ When running on the CuDNN backend, two further options must be set::
 
 .. warning::
 
-    Deterministic mode can have a performance impact, depending on your model.
+    Deterministic mode can have a performance impact, depending on your model. This means that due to the deterministic nature of the model, the processing speed (i.e. processed batch items per second) can be lower than when the model is non-deterministic.
 
 Numpy
 .....
