Fix to the Export ExecuTorch section (#506)

Summary:
Pull Request resolved: #506

* Changed the Export ExecuTorch section not to be a redirect but provide entrypoints to other sections. This provides a better UX.
* Fix a few typos.

Differential Revision: https://internalfb.com/D49706614

fbshipit-source-id: c74d94d789c11e14e1f50411ed02fa7cd66f6fd7

Author: Svetlana Karslioglu

docs/source/conf.py

@@ -68,6 +68,8 @@
     "colon_fence",
 ]
 
+myst_heading_anchors = 3
+
 sphinx_gallery_conf = {
     "examples_dirs": ["tutorials_source"],
     "gallery_dirs": ["tutorials"],

docs/source/export-overview.md

@@ -0,0 +1,15 @@
+# Exporting to Executorch
+
+One of the important steps in getting your PyTorch programs ready for execution
+on an edge device is exporting them. This is achieved through the use of a
+PyTorch API called `torch.export`.
+
+The `torch.export` documentation, which is part of the PyTorch core library, can
+be found in the Core PyTorch documentation set. Additionally, we provide a
+step-by-step tutorial that takes you through the process of exporting a PyTorch
+program, making it easier for you to understand and implement the process.
+
+To learn more about exporting your model:
+
+* Complete the [Exporting to ExecuTorch tutorial](./tutorials/export-to-executorch-tutorial).
+* Read the [torch.export documentation](https://pytorch.org/docs/2.1/export.html).

docs/source/export-overview.rst

@@ -263,6 +263,6 @@ ExecuTorch program. Now that you have a basic understanding of how ExecuTorch
 works, you can start exploring its advanced features and capabilities. Here
 is a list of sections you might want to read next:
 
-* [Exporting a model](export-overview.rst)
+* [Exporting a model](export-overview.md)
 * Using [EXIR](ir-exir.md) for advanced exports
 * Review more advanced examples in the [executorch/examples](https://github.com/pytorch/executorch/tree/main/examples) directory

docs/source/export-user-guide.rst

@@ -92,12 +92,11 @@ Topics in this section will help you get started with ExecuTorch.
    :hidden:
 
    export-overview
-   export-user-guide
 
 .. toctree::
    :glob:
    :maxdepth: 1
-   :caption: Intermediate Representation (IR) Specification
+   :caption: IR Specification
    :hidden:
 
    ir-exir

docs/source/getting-started-setup.md

@@ -15,7 +15,7 @@
 ######################################################################
 # ExecuTorch is a unified ML stack for lowering PyTorch models to edge devices.
 # It introduces improved entry points to perform model, device, and/or use-case
-# specific optizations such as backend delegation, user-defined compiler
+# specific optimizations such as backend delegation, user-defined compiler
 # transformations, default or user-defined memory planning, and more.
 #
 # At a high level, the workflow looks as follows:
@@ -62,7 +62,7 @@
 # ``torch.export``.
 #
 # Both APIs take in a model (any callable or ``torch.nn.Module``), a tuple of
-# positional arguments, optionally a dictionary of keywork arguments (not shown
+# positional arguments, optionally a dictionary of keyword arguments (not shown
 # in the example), and a list of constraints (covered later).
 
 import torch
@@ -96,7 +96,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
 # The output of ``torch._export.capture_pre_autograd_graph`` is a fully
 # flattened graph (meaning the graph does not contain any module hierarchy,
 # except in the case of control flow operators). Furthermore, the captured graph
-# contains only ATen operators (~3000 ops) which are autograd safe, i.e. safe
+# contains only ATen operators (~3000 ops) which are Autograd safe, for example, safe
 # for eager mode training.
 #
 # The output of ``torch.export`` further compiles the graph to a lower and
@@ -116,7 +116,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
 # Since the result of ``torch.export`` is a graph containing the Core ATen
 # operators, we will call this the ``ATen Dialect``, and since
 # ``torch._export.capture_pre_autograd_graph`` returns a graph containing the
-# set of ATen operators which are autograd safe, we will call it the
+# set of ATen operators which are Autograd safe, we will call it the
 # ``Pre-Autograd ATen Dialect``.
 
 ######################################################################
@@ -231,7 +231,7 @@ def f(x, y):
 # `FX Graph Mode Quantization <https://pytorch.org/tutorials/prototype/fx_graph_mode_ptq_static.html>`__,
 # we will need to call two new APIs: ``prepare_pt2e`` and ``compare_pt2e``
 # instead of ``prepare_fx`` and ``convert_fx``. It differs in that
-# ``prepare_pt2e`` takes a backend-specific ``Quantizer`` as an arugument, which
+# ``prepare_pt2e`` takes a backend-specific ``Quantizer`` as an argument, which
 # will annotate the nodes in the graph with information needed to quantize the
 # model properly for a specific backend.