update doc

thuml · Nov 27, 2023 · 8655264 · 8655264
1 parent bacdf0c
commit 8655264
Showing 1 changed file with 5 additions and 1 deletion.
diff --git a/docs/walk_through.rst b/docs/walk_through.rst
@@ -35,7 +35,11 @@ The tutorial intends to cover the following aspects of PyTorch compiler:
 - AOTAutograd (generate backward computation graph from forward computation graph)
 - Inductor/Other backends (given a computation graph, how to run it faster in different devices)
 
-These components will be called with different backend options: ``torch.compile(backend="eager")`` for only using Dynamo, ``torch.compile(backend="aot_eager")`` for using Dynamo and AOTAutograd, and ``torch.compile(backend="inductor")`` (the default argument) for using Dynamo and AOTAutograd and PyTorch's builtin graph optimization backend named ``Inductor``.
+These components will be called with different backend options:
+
+- ``torch.compile(backend="eager")`` for only using Dynamo
+- ``torch.compile(backend="aot_eager")`` for using Dynamo and AOTAutograd
+- ``torch.compile(backend="inductor")`` (the default argument) for using Dynamo and AOTAutograd and PyTorch's builtin graph optimization backend named ``Inductor``.
 
 PyTorch compiler is a Just-In-Time compiler
 --------------------------------------------