correct doc that is reported in pytorch/pytorch#113522 (#2683)

jingxu10 · svekars · web-flow · commit cad48392f062 · 2024-07-08T15:08:57.000-07:00
Co-authored-by: Svetlana Karslioglu &lt;svekars@meta.com&gt;
diff --git a/intermediate_source/scaled_dot_product_attention_tutorial.py b/intermediate_source/scaled_dot_product_attention_tutorial.py
@@ -114,7 +114,7 @@ def benchmark_torch_function_in_microseconds(f, *args, **kwargs):
 #
 # Depending on what machine you ran the above cell on and what hardware is
 # available, your results might be different.
-# - If you don’t have a GPU and are running on CPU then the context manager
+# - If you don’t have a GPU and are running on CPU then with FP32 the context manager
 # will have no effect and all three runs should return similar timings.
 # - Depending on what compute capability your graphics card supports
 # flash attention or memory efficient might have failed.

Original file line number	Diff line number	Diff line change
`@@ -114,7 +114,7 @@ def benchmark_torch_function_in_microseconds(f, args, *kwargs):`
`114`	`114`	`#`
`115`	`115`	`# Depending on what machine you ran the above cell on and what hardware is`
`116`	`116`	`# available, your results might be different.`
`117`		`-# - If you don’t have a GPU and are running on CPU then the context manager`
	`117`	`+# - If you don’t have a GPU and are running on CPU then with FP32 the context manager`
`118`	`118`	`# will have no effect and all three runs should return similar timings.`
`119`	`119`	`# - Depending on what compute capability your graphics card supports`
`120`	`120`	`# flash attention or memory efficient might have failed.`