Update

Author: Svetlana Karslioglu

distributed/home.rst

@@ -14,9 +14,9 @@ PyTorch with each method having their advantages in certain use cases:
 * `DistributedDataParallel (DDP) <#learn-ddp>`__
 * `Fully Sharded Data Parallel (FSDP) <#learn-fsdp>`__
 * `Remote Procedure Call (RPC) distributed training <#learn-rpc>`__
-* `Pipeline Parallelism <#learn-pipeline-parallelism>`__
+* `Custom Extensions <#custom-extensions>`__
 
-Read more about these options in [Distributed Overview](../beginner/dist_overview.rst).
+Read more about these options in `Distributed Overview <../beginner/dist_overview.html>`__.
 
 .. _learn-ddp:
 
@@ -47,16 +47,23 @@ Learn DDP
         +++
         :octicon:`code;1em` Code
 
+     .. grid-item-card:: :octicon:`file-code;1em`
+        Distributed Training with Uneven Inputs Using
+        the Join Context Manager
+        :shadow: none
+        :link: ../advanced_source/generic_join.rst
+        :link-type: url
+
+        This tutorial provides a short and gentle intro to the PyTorch
+        DistributedData Parallel.
+        +++
+        :octicon:`code;1em` Code
+
 .. _learn-fsdp:
 
 Learn FSDP
 ----------
 
-Fully-Sharded Data Parallel (FSDP) is a tool that distributes model
-parameters across multiple workers, therefore enabling you to train larger
-models.
-
-
 .. grid:: 3
 
      .. grid-item-card:: :octicon:`file-code;1em`
@@ -86,9 +93,6 @@ models.
 Learn RPC
 ---------
 
-Distributed Remote Procedure Call (RPC) framework provides
-mechanisms for multi-machine model training
-
 .. grid:: 3
 
      .. grid-item-card:: :octicon:`file-code;1em`
@@ -114,38 +118,44 @@ mechanisms for multi-machine model training
         :octicon:`code;1em` Code
 
      .. grid-item-card:: :octicon:`file-code;1em`
-        Distributed Pipeline Parallelism Using RPC
+        Implementing Batch RPC Processing Using Asynchronous Executions
         :shadow: none
         :link: https://example.com
         :link-type: url
 
-        Learn how to use a Resnet50 model for distributed pipeline parallelism
-        with the Distributed RPC APIs.
+        In this tutorial you will build batch-processing RPC applications
+        with the @rpc.functions.async_execution decorator.
         +++
         :octicon:`code;1em` Code
 
 .. grid:: 3
 
      .. grid-item-card:: :octicon:`file-code;1em`
-        Implementing Batch RPC Processing Using Asynchronous Executions
+        Combining Distributed DataParallel with Distributed RPC Framework
         :shadow: none
         :link: https://example.com
         :link-type: url
 
-        In this tutorial you will build batch-processing RPC applications
-        with the @rpc.functions.async_execution decorator.
+        In this tutorial you will learn how to combine distributed data
+        parallelism with distributed model parallelism.
         +++
         :octicon:`code;1em` Code
 
+.. _custom-extensions:
+
+Custom Extensions
+-----------------
+
+.. grid:: 3
+
      .. grid-item-card:: :octicon:`file-code;1em`
-        Combining Distributed DataParallel with Distributed RPC Framework
+        Customize Process Group Backends Using Cpp Extensions
         :shadow: none
-        :link: https://example.com
+        :link: intermediate/process_group_cpp_extension_tutorial.html
         :link-type: url
 
-        In this tutorial you will learn how to combine distributed data
-        parallelism with distributed model parallelism.
+        In this tutorial you will learn to implement a custom `ProcessGroup`
+        backend and plug that into PyTorch distributed package using
+        cpp extensions.
         +++
         :octicon:`code;1em` Code
-
-.. _learn-pipeline-parallelism: